[ { "number": 3223, "title": "TorchaoLoraLinear: __init__() missing get_apply_tensor_subclass kwarg with torchao 0.17 (PEFT 0.19.1)", "state": "open", "author": "genno-whittlery", "created_at": "2026-05-11T13:22:31Z", "updated_at": "2026-05-12T10:41:28Z", "html_url": "https://github.com/huggingface/peft/issues/3223", "labels": [], "model": "FLUX.2", "trainer": "peft_direct", "specialties": [ "torchao", "quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Comprehensive bug report with minimal reproducer, full environment table, traceback, and workaround.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3229, "title": "[BUG] peft 0.19 target_modules (str) use `set`", "state": "open", "author": "Jintao-Huang", "created_at": "2026-05-12T10:17:08Z", "updated_at": "2026-05-12T10:17:08Z", "html_url": "https://github.com/huggingface/peft/issues/3229", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 6, "experience_rationale": "Points to exact source lines and references external issue, but no runnable reproduction or environment details.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "medium", "peft_method": "low" } }, { "number": 2310, "title": "Comparison of Different Fine-Tuning Techniques for Conversational AI", "state": "open", "author": "ImamaDev", "created_at": "2025-01-07T07:07:50Z", "updated_at": "2026-05-12T09:48:54Z", "html_url": "https://github.com/huggingface/peft/issues/2310", "labels": [ "good first issue", "help wanted", "contributions-welcome" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 4, "experience_rationale": "Well-intentioned feature request with list of methods but lacks technical depth, reproducible example, or specific use case.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low" } }, { "number": 3129, "title": "Add support for Gemma4ClippableLinear (Gemma 4 QLoRA fails)", "state": "closed", "author": "dentity007", "created_at": "2026-04-02T20:15:42Z", "updated_at": "2026-05-11T08:48:32Z", "html_url": "https://github.com/huggingface/peft/issues/3129", "labels": [], "model": "Gemma", "trainer": "peft_direct", "specialties": [ "bitsandbytes", "quantization", "4bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Detailed bug report with root cause analysis, minimal workaround example, and clear error message.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3142, "title": "Proposal: add MiCA (Minor Component Adaptation) as a PEFT method", "state": "open", "author": "sr-networks", "created_at": "2026-04-09T09:11:49Z", "updated_at": "2026-05-10T15:23:41Z", "html_url": "https://github.com/huggingface/peft/issues/3142", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "MiCA", "experience_score": 7, "experience_rationale": "Well-structured feature proposal with paper reference and implementation plan, but no code or empirical evidence provided.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3123, "title": "Nemotron-3-Nano NemotronHTopkRouter", "state": "closed", "author": "lucian-student", "created_at": "2026-03-30T10:10:19Z", "updated_at": "2026-05-10T15:21:58Z", "html_url": "https://github.com/huggingface/peft/issues/3123", "labels": [], "model": "Nemotron", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Includes layer implementation code and clear request, but lacks reproducible example or training context.", "confidence": { "model": "high", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2944, "title": "enusre_weight_tying = True does not respect the model config tie_word_embeddings flag", "state": "closed", "author": "romitjain", "created_at": "2025-12-09T08:36:16Z", "updated_at": "2026-05-09T08:10:24Z", "html_url": "https://github.com/huggingface/peft/issues/2944", "labels": [], "model": "Mistral", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Provides runnable minimal reproduction with assertions and specific source code reference.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3182, "title": "RFC: Improve code to resolve LoRA variants", "state": "open", "author": "BenjaminBossan", "created_at": "2026-04-21T12:59:01Z", "updated_at": "2026-05-08T13:36:31Z", "html_url": "https://github.com/huggingface/peft/issues/3182", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Technical RFC with code references and concrete proposal, but no empirical data or reproduction.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3114, "title": "GraLoRA merge_and_unload with modules_to_save error", "state": "open", "author": "JBurtn", "created_at": "2026-03-22T23:25:22Z", "updated_at": "2026-05-07T15:46:45Z", "html_url": "https://github.com/huggingface/peft/issues/3114", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "GraLoRA", "experience_score": 7, "experience_rationale": "Includes full traceback and version info, but lacks minimal reproduction script and model details.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3201, "title": "PrefixTuningConfig fails on google/gemma-4-e2b-it: tensor expand size mismatch in attention forward (peft 0.19.1, transformers 5.6.2)", "state": "closed", "author": "stharrold", "created_at": "2026-04-28T11:31:39Z", "updated_at": "2026-05-06T09:43:58Z", "html_url": "https://github.com/huggingface/peft/issues/3201", "labels": [], "model": "Gemma", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "Prefix Tuning", "experience_score": 9, "experience_rationale": "Excellent bug report with complete minimal reproducer, error analysis, and cross-references to similar fixes.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3214, "title": "[Windows] RTX 5070 Ti (Blackwell sm_120) - setup and deployment notes", "state": "closed", "author": "loongmiaow-pixel", "created_at": "2026-05-05T03:54:26Z", "updated_at": "2026-05-06T09:36:27Z", "html_url": "https://github.com/huggingface/peft/issues/3214", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "bitsandbytes" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 5, "experience_rationale": "Provides hardware environment but is a general support question without reproduction or error logs.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "low", "peft_method": "medium" } }, { "number": 3211, "title": "Feature Request: Improve offline support for custom architectures in get_peft_model_state_dict", "state": "closed", "author": "nu11secur1ty", "created_at": "2026-05-03T05:38:47Z", "updated_at": "2026-05-03T12:23:46Z", "html_url": "https://github.com/huggingface/peft/issues/3211", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 5, "experience_rationale": "Clear feature request with proposed logic snippet but no reproduction or concrete model example.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low" } }, { "number": 3193, "title": "Add torch.nn.Conv2d support in AdaLoRA", "state": "closed", "author": "PARRYH0TTEr", "created_at": "2026-04-24T08:03:19Z", "updated_at": "2026-04-30T15:16:30Z", "html_url": "https://github.com/huggingface/peft/issues/3193", "labels": [], "model": "ResNet", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "AdaLoRA", "experience_score": 8, "experience_rationale": "Provides minimal code example showing LoRA vs AdaLoRA difference on ResNet-18 with clear motivation.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3186, "title": "Proposal: add BEFT as a new PEFT method", "state": "closed", "author": "whubaichuan", "created_at": "2026-04-23T07:01:12Z", "updated_at": "2026-04-30T12:42:11Z", "html_url": "https://github.com/huggingface/peft/issues/3186", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "BEFT", "experience_score": 7, "experience_rationale": "Well-written proposal with paper reference and author credentials, but no code or empirical integration test provided.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3203, "title": "Proposal: FIM-guided adaptive LoRA rank allocation (FimConfig + initialize_lora_fim_ranks)", "state": "open", "author": "ramkrishs", "created_at": "2026-04-28T22:01:49Z", "updated_at": "2026-04-28T22:01:49Z", "html_url": "https://github.com/huggingface/peft/issues/3203", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Comprehensive proposal with algorithm, API design, and mathematical formulation, but no implementation or reproduction.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2864, "title": "Deal with weight tying consistently", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-10-23T13:30:39Z", "updated_at": "2026-04-28T15:57:36Z", "html_url": "https://github.com/huggingface/peft/issues/2864", "labels": [ "contributions-welcome" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Technical design discussion with detailed behavior matrix and references to prior work, but no runnable code or model-specific context.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3191, "title": "AuxiliaryTrainingWrapper.forward requires positional `x`, breaks kwargs-only modules in modules_to_save", "state": "closed", "author": "ved1beta", "created_at": "2026-04-23T18:51:32Z", "updated_at": "2026-04-26T19:03:26Z", "html_url": "https://github.com/huggingface/peft/issues/3191", "labels": [], "model": "Gemma", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Excellent bug report with minimal standalone reproducer and real-world Gemma 4 VLM example with YAML config.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3004, "title": "Combining adapters linearly with negative weights is broken", "state": "closed", "author": "orr-hirundo", "created_at": "2026-01-21T11:25:13Z", "updated_at": "2026-04-26T15:16:35Z", "html_url": "https://github.com/huggingface/peft/issues/3004", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Technical bug report referencing exact source lines and explaining mathematical issue, but lacks runnable reproduction.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3044, "title": "TP support for Finetuning using LoRA and other PEFT techniques", "state": "closed", "author": "quic-meetkuma", "created_at": "2026-02-16T08:34:36Z", "updated_at": "2026-04-23T15:42:25Z", "html_url": "https://github.com/huggingface/peft/issues/3044", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Clear feature request with technical reasoning about tensor parallelism, but no concrete reproduction or model example.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3184, "title": "Proposal: support ShadowPEFT method in PEFT", "state": "open", "author": "SeanLee97", "created_at": "2026-04-22T04:43:52Z", "updated_at": "2026-04-22T09:36:51Z", "html_url": "https://github.com/huggingface/peft/issues/3184", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "ShadowPEFT", "experience_score": 7, "experience_rationale": "Well-structured proposal with paper and model collection links, but no code or training example provided.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3080, "title": "FSDP + torch.nn.Parameter (MoE layer) lora fine-tuning doesn't work", "state": "closed", "author": "dskhudia", "created_at": "2026-03-04T06:39:47Z", "updated_at": "2026-04-20T15:30:53Z", "html_url": "https://github.com/huggingface/peft/issues/3080", "labels": [], "model": "GPT-OSS", "trainer": "trl", "specialties": [ "FSDP" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Includes full error traceback, reproduction script with training setup, and environment details (8x H100).", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3169, "title": "LoRA + BnB INT8 + CPU offload: output tensor on wrong device in tuners/lora/bnb.py", "state": "open", "author": "sirfyyn", "created_at": "2026-04-16T15:57:38Z", "updated_at": "2026-04-20T10:14:24Z", "html_url": "https://github.com/huggingface/peft/issues/3169", "labels": [], "model": "Gemma", "trainer": "peft_direct", "specialties": [ "bitsandbytes", "quantization", "8bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Detailed bug report with specific fix locations, training context, and link to complete training example patches.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3085, "title": "Security: Arbitrary Module Import via Malicious Adapter Config (CWE-94)", "state": "closed", "author": "spartan8806", "created_at": "2026-03-08T22:51:09Z", "updated_at": "2026-04-16T16:30:07Z", "html_url": "https://github.com/huggingface/peft/issues/3085", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 9, "experience_rationale": "Professional security report with CWE classification, detailed attack vectors, and minimal reproduction steps.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low" } }, { "number": 3158, "title": "AttributeError: module 'torch' has no attribute 'float8_e8m0fnu' when casting adapter dtype on PyTorch 2.5.1", "state": "closed", "author": "yoogie27", "created_at": "2026-04-14T19:12:35Z", "updated_at": "2026-04-15T15:38:46Z", "html_url": "https://github.com/huggingface/peft/issues/3158", "labels": [], "model": "unknown", "trainer": "unsloth", "specialties": [ "bitsandbytes", "quantization", "4bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Clear bug report with traceback, temporary workaround, and bash reproduction script for environment setup.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3020, "title": "LoRA + activation checkpointing breaks model compilation with torch.compile when dropout != 0", "state": "closed", "author": "raliouat", "created_at": "2026-02-03T08:59:09Z", "updated_at": "2026-04-14T15:27:28Z", "html_url": "https://github.com/huggingface/peft/issues/3020", "labels": [], "model": "LLaVA", "trainer": "accelerate", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Good bug report with minimal reproducer, clear condition (dropout != 0), and specific model mentioned.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 2970, "title": "Add EWoRA to PEFT", "state": "closed", "author": "HarshKohli", "created_at": "2025-12-31T08:47:34Z", "updated_at": "2026-04-12T15:14:06Z", "html_url": "https://github.com/huggingface/peft/issues/2970", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "EWoRA", "experience_score": 5, "experience_rationale": "Brief feature request with paper reference but lacks detail on implementation or model specifics.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3134, "title": "Inquiry About Two-Stage QLoRA Fine-Tuning", "state": "closed", "author": "AAnirudh07", "created_at": "2026-04-07T04:19:32Z", "updated_at": "2026-04-11T05:55:26Z", "html_url": "https://github.com/huggingface/peft/issues/3134", "labels": [], "model": "Qwen", "trainer": "peft_direct", "specialties": [ "bitsandbytes", "quantization", "4bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Well-structured question about two-stage QLoRA pipeline with specific model and clear concerns about degradation.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3141, "title": "Proposal: add MiCA (Minor Component Adaptation) as a new PEFT method", "state": "closed", "author": "rankofootball", "created_at": "2026-04-09T09:06:14Z", "updated_at": "2026-04-09T09:08:47Z", "html_url": "https://github.com/huggingface/peft/issues/3141", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "MiCA", "experience_score": 7, "experience_rationale": "Duplicate-ish proposal of MiCA with paper reference and implementation sketch, but no code or model example.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3130, "title": "Add LoRA support for Gemma 4 (Gemma4ClippableLinear)", "state": "closed", "author": "skap3214", "created_at": "2026-04-02T22:08:20Z", "updated_at": "2026-04-03T11:15:54Z", "html_url": "https://github.com/huggingface/peft/issues/3130", "labels": [], "model": "Gemma", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Comprehensive bug report with root cause analysis, class implementation, and multiple workarounds.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 2945, "title": "Return base model state_dict with original keys", "state": "open", "author": "dvmazur", "created_at": "2025-12-09T11:23:52Z", "updated_at": "2026-04-01T11:20:46Z", "html_url": "https://github.com/huggingface/peft/issues/2945", "labels": [ "wip" ], "model": "unknown", "trainer": "unknown", "specialties": [ "FSDP" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Clear feature request with technical context about FSDP and state dict key handling, but no reproduction or model example.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "medium" } }, { "number": 1988, "title": "TypeError: WhisperForConditionalGeneration.forward() got an unexpected keyword argument 'input_ids'", "state": "closed", "author": "YOUSEFNANIS", "created_at": "2024-08-02T17:48:46Z", "updated_at": "2026-03-26T12:12:20Z", "html_url": "https://github.com/huggingface/peft/issues/1988", "labels": [], "model": "Whisper", "trainer": "peft_direct", "specialties": [ "bitsandbytes", "quantization", "8bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Includes runnable reproduction with specific model and versions, but lacks clear expected behavior description.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3116, "title": "Evaluation protocol uses best validation score as reported metric and inflates results", "state": "closed", "author": "paolomandica", "created_at": "2026-03-25T13:46:20Z", "updated_at": "2026-03-25T15:15:25Z", "html_url": "https://github.com/huggingface/peft/issues/3116", "labels": [], "model": "RoBERTa", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Thoughtful methodology critique with reference to original paper and personal experiment results, but no training code or model configuration shown.", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "medium" } }, { "number": 3043, "title": "trainable_token_indices of LoraConfig not working when using more than 1 NVIDIA GPU", "state": "closed", "author": "Oidlichtnwoada", "created_at": "2026-02-14T15:05:42Z", "updated_at": "2026-03-24T15:26:09Z", "html_url": "https://github.com/huggingface/peft/issues/3043", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Includes partial traceback showing device mismatch error, but lacks minimal reproduction and model details.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2916, "title": "ParamWrapper initialization fails when DeepSpeed ZeRO-3 is enabled", "state": "closed", "author": "jonnyli1125", "created_at": "2025-11-18T21:32:30Z", "updated_at": "2026-03-24T11:23:08Z", "html_url": "https://github.com/huggingface/peft/issues/2916", "labels": [], "model": "GPT", "trainer": "trl", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Good bug report with minimal reproduction using TRL SFTConfig and DeepSpeed ZeRO-3, plus exact source line reference.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3033, "title": "Create Baseline Evaluation Based on DriveLM data off-the-shelf VLMs", "state": "closed", "author": "Hamidreza3252", "created_at": "2026-02-09T22:51:15Z", "updated_at": "2026-03-21T15:06:50Z", "html_url": "https://github.com/huggingface/peft/issues/3033", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 2, "experience_rationale": "Extremely vague feature request with no technical details, models, or reproduction.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low" } }, { "number": 2861, "title": "[FEAT] Integrate WoRA (Filtering-WoRA) into PEFT", "state": "closed", "author": "JT-Sun", "created_at": "2025-10-23T02:36:20Z", "updated_at": "2026-03-21T08:46:08Z", "html_url": "https://github.com/huggingface/peft/issues/2861", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "WoRA", "experience_score": 7, "experience_rationale": "Well-structured feature proposal with paper reference and performance figures, but no code or model-specific details.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3110, "title": "Discrepancy between layers names from `get_peft_model` and `.load_adapter`", "state": "closed", "author": "AngledLuffa", "created_at": "2026-03-17T18:54:05Z", "updated_at": "2026-03-18T14:58:09Z", "html_url": "https://github.com/huggingface/peft/issues/3110", "labels": [], "model": "ELECTRA", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Includes configuration output and parameter name listings demonstrating the naming discrepancy, but no minimal standalone reproducer.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3002, "title": "Running Lora+ T5 Gemma 2 on multi gpu results in error", "state": "closed", "author": "silviiaromanato", "created_at": "2026-01-19T15:21:27Z", "updated_at": "2026-03-17T15:24:33Z", "html_url": "https://github.com/huggingface/peft/issues/3002", "labels": [], "model": "T5", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Includes full error message, environment specs, and specific model, but no minimal reproduction script.", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 3012, "title": "Question about multiple LoRAs and FSDP", "state": "closed", "author": "l1346792580123", "created_at": "2026-01-28T02:19:46Z", "updated_at": "2026-03-17T15:24:29Z", "html_url": "https://github.com/huggingface/peft/issues/3012", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "FSDP" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 5, "experience_rationale": "Clear architectural question with pseudocode, but lacks model details, environment, or reproduction.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2998, "title": "Tutorial uses deprecated dataset (financial_phrasebank) that fails to load", "state": "closed", "author": "maerory", "created_at": "2026-01-16T09:52:21Z", "updated_at": "2026-03-17T11:21:48Z", "html_url": "https://github.com/huggingface/peft/issues/2998", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "IA3", "experience_score": 6, "experience_rationale": "Clear documentation bug report with specific tutorial link and reproduction code, but no model or training context.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2882, "title": "[FEAT] Integrate LoRA-One into PEFT", "state": "closed", "author": "YuanheZ", "created_at": "2025-10-30T16:57:49Z", "updated_at": "2026-03-16T15:25:36Z", "html_url": "https://github.com/huggingface/peft/issues/2882", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA-One", "experience_score": 7, "experience_rationale": "Well-structured feature proposal with paper reference and detailed contribution list, but no code or model example.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3089, "title": "Can LoRA ignore standard modules it doesn't know about?", "state": "closed", "author": "AngledLuffa", "created_at": "2026-03-10T19:11:14Z", "updated_at": "2026-03-16T12:07:37Z", "html_url": "https://github.com/huggingface/peft/issues/3089", "labels": [], "model": "ELECTRA", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Includes actual error message and desired behavior, but no reproduction script or environment details.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "medium", "peft_method": "high" } }, { "number": 2930, "title": "[FEAT] Integrate FeRA: Frequency-Energy Constrained Routing for Effective Diffusion Adaptation Fine-Tuning", "state": "closed", "author": "YinBo0927", "created_at": "2025-11-29T03:09:28Z", "updated_at": "2026-03-15T15:08:48Z", "html_url": "https://github.com/huggingface/peft/issues/2930", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "FeRA", "experience_score": 6, "experience_rationale": "Feature proposal with paper reference and visual diagrams, but lacks implementation details or model specifics.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3016, "title": "layers_to_transform incorrectly matches MoE expert indices (should match layer index)", "state": "closed", "author": "MichalMraz", "created_at": "2026-02-01T11:47:15Z", "updated_at": "2026-03-12T15:19:50Z", "html_url": "https://github.com/huggingface/peft/issues/3016", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Excellent bug report with minimal toy model reproduction, clear expected vs actual behavior, and proposed fix.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2941, "title": "Add MonteCLoRA (TMLR 2025 accepted) in PEFT library", "state": "closed", "author": "victor7246", "created_at": "2025-12-07T08:45:21Z", "updated_at": "2026-03-08T15:05:59Z", "html_url": "https://github.com/huggingface/peft/issues/2941", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "MonteCLoRA", "experience_score": 4, "experience_rationale": "Very brief feature request with paper link but no details on implementation or model specifics.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2563, "title": "Integrate Lily", "state": "closed", "author": "yibozhong", "created_at": "2025-06-02T07:23:30Z", "updated_at": "2026-03-04T13:08:52Z", "html_url": "https://github.com/huggingface/peft/issues/2563", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "Lily", "experience_score": 7, "experience_rationale": "Well-structured proposal with paper and repository links, detailed motivation, and discussion of design variants.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2928, "title": "Inference mode with Module_to_save LoRA", "state": "closed", "author": "NguyenRichard", "created_at": "2025-11-27T21:56:53Z", "updated_at": "2026-03-04T10:46:02Z", "html_url": "https://github.com/huggingface/peft/issues/2928", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [ "bitsandbytes", "quantization", "8bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Includes custom wrapper code and inference context, but reproduction is complex and model details are missing.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "medium", "peft_method": "high" } }, { "number": 3073, "title": "LoRA gradients not normalized by input norm → training instability (NaN)", "state": "closed", "author": "warjohn", "created_at": "2026-03-01T18:59:20Z", "updated_at": "2026-03-03T18:01:29Z", "html_url": "https://github.com/huggingface/peft/issues/3073", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Technical report with mathematical analysis of gradient scaling and provided workaround hook code, but no minimal reproduction.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "medium", "peft_method": "high" } }, { "number": 2948, "title": "Add PVeRA to PEFT", "state": "closed", "author": "leofillioux", "created_at": "2025-12-10T16:38:42Z", "updated_at": "2026-03-03T10:07:13Z", "html_url": "https://github.com/huggingface/peft/issues/2948", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "PVeRA", "experience_score": 5, "experience_rationale": "Brief feature request with paper reference and accuracy numbers, but no implementation details or model examples.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3026, "title": "Add PSOFT to PEFT", "state": "closed", "author": "fei407", "created_at": "2026-02-07T10:20:17Z", "updated_at": "2026-03-02T21:49:09Z", "html_url": "https://github.com/huggingface/peft/issues/3026", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "PSOFT", "experience_score": 6, "experience_rationale": "Feature proposal with paper reference and theoretical motivation, but no code or concrete model example provided.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3075, "title": "Error When Loading Private Adapters", "state": "closed", "author": "TyTodd", "created_at": "2026-03-02T07:02:56Z", "updated_at": "2026-03-02T07:06:45Z", "html_url": "https://github.com/huggingface/peft/issues/3075", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 6, "experience_rationale": "Includes error logs showing 401 Unauthorized, but no minimal reproduction or environment details beyond basic stack trace.", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low" } }, { "number": 3025, "title": "Detail: removal of `.peft_config` when performing `.unload()`?", "state": "closed", "author": "jchwenger", "created_at": "2026-02-06T20:13:50Z", "updated_at": "2026-02-27T14:02:56Z", "html_url": "https://github.com/huggingface/peft/issues/3025", "labels": [], "model": "Gemma", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Clear question with code snippets demonstrating the issue, but no error logs or environment details.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3049, "title": "More efficient PISSA initialization for LoRA (faster/cheaper than current SVD)", "state": "closed", "author": "Brokttv", "created_at": "2026-02-18T08:59:30Z", "updated_at": "2026-02-26T10:13:20Z", "html_url": "https://github.com/huggingface/peft/issues/3049", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 5, "experience_rationale": "Brief feature inquiry about PISSA optimization with mention of Krylov method, but no concrete proposal or reproduction.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2977, "title": "Unable to use enable/disable_adapter_layers in FSDP training", "state": "closed", "author": "varunlmxd", "created_at": "2026-01-07T19:21:07Z", "updated_at": "2026-02-19T12:36:17Z", "html_url": "https://github.com/huggingface/peft/issues/2977", "labels": [], "model": "Mistral", "trainer": "peft_direct", "specialties": [ "FSDP" ], "training_type": "supervised", "peft_method": "unknown", "experience_score": 5, "experience_rationale": "Includes code snippet showing the error pattern but lacks minimal reproduction, model details, and environment information.", "confidence": { "model": "high", "trainer_framework": "medium", "specialties": "high", "training_type": "medium", "peft_method": "low" } }, { "number": 3039, "title": "extremely slow get_peft_model() with LoRA", "state": "closed", "author": "EaswarGn", "created_at": "2026-02-12T01:42:30Z", "updated_at": "2026-02-18T10:22:52Z", "html_url": "https://github.com/huggingface/peft/issues/3039", "labels": [], "model": "Qwen", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Includes reproduction code showing the performance issue with specific model, but no profiling data or environment details.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 871, "title": "unnecessarily slow get_peft_model() with LoRA (esp. in LLMs)", "state": "closed", "author": "poedator", "created_at": "2023-08-28T08:41:33Z", "updated_at": "2026-02-13T06:34:33Z", "html_url": "https://github.com/huggingface/peft/issues/871", "labels": [ "enhancement" ], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Includes minimal benchmark script with timing measurement, but uses synthetic model rather than real LLM.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2813, "title": "[FEAT] Integrate LoRETTA: Low-Rank Economic Tensor-Train Adaptation for Ultra-Low-Parameter Fine-Tuning of Large Language Models", "state": "closed", "author": "mbaddar1", "created_at": "2025-10-06T20:53:55Z", "updated_at": "2026-02-13T05:59:36Z", "html_url": "https://github.com/huggingface/peft/issues/2813", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRETTA", "experience_score": 6, "experience_rationale": "Feature proposal with paper reference and performance comparison figures, but no code or model-specific details beyond BERT mention.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3034, "title": "Evaluate augmenting driveLM data with spatial SG using off-the-shelf VLM", "state": "closed", "author": "Hamidreza3252", "created_at": "2026-02-09T23:13:17Z", "updated_at": "2026-02-12T11:15:30Z", "html_url": "https://github.com/huggingface/peft/issues/3034", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 2, "experience_rationale": "Extremely vague feature request with no technical details, models, or reproduction.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low" } }, { "number": 2949, "title": "Deprecation of warmup_ratio argument in transformers v5", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-12-11T11:21:41Z", "updated_at": "2026-02-12T11:10:23Z", "html_url": "https://github.com/huggingface/peft/issues/2949", "labels": [ "contributions-welcome" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 5, "experience_rationale": "Clear documentation/dependency issue with reference to upstream PR, but no model or training context.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low" } }, { "number": 2973, "title": "Add BoRA to PEFT", "state": "closed", "author": "cxmscb", "created_at": "2026-01-04T03:25:56Z", "updated_at": "2026-02-11T15:21:46Z", "html_url": "https://github.com/huggingface/peft/issues/2973", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "BoRA", "experience_score": 4, "experience_rationale": "Very brief feature request with paper link and image, lacking implementation details or model specifics.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 1384, "title": "Saving and Loading fine-tuned model", "state": "closed", "author": "binbinxue", "created_at": "2024-01-22T14:14:40Z", "updated_at": "2026-02-07T10:48:21Z", "html_url": "https://github.com/huggingface/peft/issues/1384", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "bitsandbytes", "quantization", "4bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 5, "experience_rationale": "Conceptual question about quantization consistency during inference with saved adapters, but no reproduction or model details.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2342, "title": "CI: Add gptqmodel to the CI", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-01-23T12:57:29Z", "updated_at": "2026-01-29T13:33:00Z", "html_url": "https://github.com/huggingface/peft/issues/2342", "labels": [ "wip" ], "model": "unknown", "trainer": "unknown", "specialties": [ "GPTQ" ], "training_type": "none", "peft_method": "unknown", "experience_score": 4, "experience_rationale": "Brief CI tracking issue with reference to PR comment, no reproduction or model context.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "high", "peft_method": "low" } }, { "number": 1580, "title": "PeftModel is_trainable=True causes generate output to be garbage.", "state": "closed", "author": "o1lo01ol1o", "created_at": "2024-03-21T12:49:53Z", "updated_at": "2026-01-29T11:43:00Z", "html_url": "https://github.com/huggingface/peft/issues/1580", "labels": [], "model": "Mistral", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Includes runnable reproduction with specific model and clear example of garbage output, but environment details are missing.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 2907, "title": "Integrate: LoRA vs Full Fine-tuning: An Illusion of Equivalence", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-11-12T15:49:11Z", "updated_at": "2026-01-28T00:35:32Z", "html_url": "https://github.com/huggingface/peft/issues/2907", "labels": [ "contributions-welcome" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Well-structured feature request with paper reference and detailed API design sketch, but no model or empirical data.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 3009, "title": "peft failed to support LoRA for MoE models with fused expert weights in transformers v5", "state": "closed", "author": "ITcarrot", "created_at": "2026-01-24T19:10:16Z", "updated_at": "2026-01-27T16:06:22Z", "html_url": "https://github.com/huggingface/peft/issues/3009", "labels": [], "model": "Qwen", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Good bug report with reproduction code showing the MoE implementation change and its impact on PEFT LoRA dispatch.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 1340, "title": "Training a PeftModel loaded with `from_pretrained` causes a grad-related RuntimeError", "state": "closed", "author": "EricLBuehler", "created_at": "2024-01-10T02:00:03Z", "updated_at": "2026-01-22T04:13:51Z", "html_url": "https://github.com/huggingface/peft/issues/1340", "labels": [], "model": "Mistral", "trainer": "transformers.Trainer", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Includes simplified reproduction code and error message, but lacks minimal standalone script and environment details.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 2927, "title": "Add LoRA-GA (Low-Rank Adaptation with Gradient Approximation) to PEFT", "state": "closed", "author": "sambhavnoobcoder", "created_at": "2025-11-26T21:28:04Z", "updated_at": "2026-01-15T19:05:10Z", "html_url": "https://github.com/huggingface/peft/issues/2927", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA-GA", "experience_score": 7, "experience_rationale": "Well-structured feature proposal with paper reference and clear algorithm steps, but no code or model example.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2648, "title": "Feature Request: Add Adaptive Singular Value Decomposition based Orthogonal Subspace Fine-Tuning", "state": "closed", "author": "NikhilNayak-debug", "created_at": "2025-07-15T15:29:49Z", "updated_at": "2026-01-15T10:50:30Z", "html_url": "https://github.com/huggingface/peft/issues/2648", "labels": [ "wip" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 7, "experience_rationale": "Detailed mathematical proposal with memory complexity analysis, but no named PEFT method or model example provided.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low" } }, { "number": 2980, "title": "TestOpt4bitBnb::test_lora_4bit result mismatch", "state": "closed", "author": "jiqing-feng", "created_at": "2026-01-08T08:23:35Z", "updated_at": "2026-01-14T10:47:00Z", "html_url": "https://github.com/huggingface/peft/issues/2980", "labels": [], "model": "OPT", "trainer": "unknown", "specialties": [ "bitsandbytes", "quantization", "4bit_quantization" ], "training_type": "none", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Regression test failure report with exact assertion error, but no model training context or environment details beyond A100 mention.", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 2988, "title": "Evaluation code uses argmax instead of generate in prompt tuning tutorial - Prompt-based methods", "state": "closed", "author": "maerory", "created_at": "2026-01-11T08:53:40Z", "updated_at": "2026-01-14T10:20:11Z", "html_url": "https://github.com/huggingface/peft/issues/2988", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "Prompt Tuning", "experience_score": 7, "experience_rationale": "Clear documentation bug report with incorrect code snippet, proposed fix, and impact assessment, but no model-specific context.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2993, "title": "Unable to load the saved Lora weights", "state": "closed", "author": "botany233", "created_at": "2026-01-13T14:01:15Z", "updated_at": "2026-01-14T05:35:01Z", "html_url": "https://github.com/huggingface/peft/issues/2993", "labels": [], "model": "Qwen", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Includes save/load code snippets and version information, but lacks error traceback and minimal standalone reproduction.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 2991, "title": "CoRDA initialization lacks support for Conv1D layers in older models like GPT-2", "state": "closed", "author": "sambhavnoobcoder", "created_at": "2026-01-13T09:37:05Z", "updated_at": "2026-01-13T17:27:56Z", "html_url": "https://github.com/huggingface/peft/issues/2991", "labels": [], "model": "GPT", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Excellent bug report with labeled reproducer script, clear expected behavior, and comparison to working EVA initialization.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 2940, "title": "[Feature Request] Auto-generate 'rank_pattern' using circuit-level sensitivity analysis", "state": "closed", "author": "songeunseo", "created_at": "2025-12-06T16:27:16Z", "updated_at": "2026-01-13T12:02:05Z", "html_url": "https://github.com/huggingface/peft/issues/2940", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Feature request with paper reference and clear differentiation from AdaLoRA, but no code or model example.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 1442, "title": "RuntimeError when combining FSDP and disable_adapter", "state": "closed", "author": "wanghao14", "created_at": "2024-02-06T15:40:21Z", "updated_at": "2026-01-13T02:11:12Z", "html_url": "https://github.com/huggingface/peft/issues/1442", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [ "FSDP" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Includes partial traceback and custom model context, but lacks minimal reproduction and environment details are buried in screenshot.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "medium", "peft_method": "high" } }, { "number": 1700, "title": "Error while loading PEFT lora model", "state": "closed", "author": "Zuhashaik", "created_at": "2024-04-30T20:36:17Z", "updated_at": "2026-01-12T02:38:03Z", "html_url": "https://github.com/huggingface/peft/issues/1700", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [ "bitsandbytes" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Includes training configuration and loading code, but error message is truncated and base model is not clearly identified.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 2975, "title": "Commit 30a19a0 (gradient checkpointing) requires transformers>=4.52.0", "state": "closed", "author": "bvolkmer", "created_at": "2026-01-06T13:20:25Z", "updated_at": "2026-01-08T15:19:42Z", "html_url": "https://github.com/huggingface/peft/issues/2975", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 6, "experience_rationale": "Clear dependency issue with specific commit reference and import traceback, but no model or training context.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low" } }, { "number": 2929, "title": "Supported `transformers` versions are not clear", "state": "closed", "author": "abdulfatir", "created_at": "2025-11-27T22:53:57Z", "updated_at": "2026-01-05T15:05:51Z", "html_url": "https://github.com/huggingface/peft/issues/2929", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 4, "experience_rationale": "Brief complaint about missing version constraints with external reference, but no concrete reproduction or model context.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low" } }, { "number": 2651, "title": "DoRA slow forward inference", "state": "closed", "author": "phemw", "created_at": "2025-07-17T04:08:29Z", "updated_at": "2026-01-02T15:05:04Z", "html_url": "https://github.com/huggingface/peft/issues/2651", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "DoRA", "experience_score": 6, "experience_rationale": "Clear performance issue report with specific code reference and expected behavior, but no reproduction or model details.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "medium", "peft_method": "high" } }, { "number": 2923, "title": "timm: ConvNormAct", "state": "closed", "author": "lucian-student", "created_at": "2025-11-20T23:18:31Z", "updated_at": "2025-12-29T15:05:09Z", "html_url": "https://github.com/huggingface/peft/issues/2923", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "Very brief feature request with no technical details or model specifics beyond timm mention.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2898, "title": "LORA adapter with new embedding tokens has size and load issue", "state": "closed", "author": "imohitmayank", "created_at": "2025-11-05T15:22:43Z", "updated_at": "2025-12-20T15:03:57Z", "html_url": "https://github.com/huggingface/peft/issues/2898", "labels": [], "model": "Gemma", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Includes detailed size comparison table and link to testing gist, but reproduction is complex and lacks clear error traceback.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 2964, "title": "Regex support for elements in target_modules list", "state": "closed", "author": "juzhxng", "created_at": "2025-12-18T13:24:33Z", "updated_at": "2025-12-19T10:59:20Z", "html_url": "https://github.com/huggingface/peft/issues/2964", "labels": [], "model": "Whisper", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Clear feature request with concrete working and non-working code examples, but no reproduction or model training context.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2889, "title": "`autocast_adapter_dtype=False` doesn't work when the model is quantized", "state": "closed", "author": "qgallouedec", "created_at": "2025-11-04T01:25:45Z", "updated_at": "2025-12-17T17:25:19Z", "html_url": "https://github.com/huggingface/peft/issues/2889", "labels": [], "model": "Qwen", "trainer": "peft_direct", "specialties": [ "bitsandbytes", "quantization", "4bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Excellent bug report with side-by-side minimal reproducer for quantized vs non-quantized behavior and clear expected output.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2957, "title": "torch compiled model cannot have LoRA added without workaround", "state": "closed", "author": "bghira", "created_at": "2025-12-15T16:47:59Z", "updated_at": "2025-12-17T11:56:21Z", "html_url": "https://github.com/huggingface/peft/issues/2957", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Excellent bug report with minimal toy model reproducer, clear error explanation, and workaround demonstration.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2958, "title": "peft assumes that torch has distributed support without checking", "state": "closed", "author": "vladmandic", "created_at": "2025-12-15T17:26:47Z", "updated_at": "2025-12-17T11:35:09Z", "html_url": "https://github.com/huggingface/peft/issues/2958", "labels": [], "model": "unknown", "trainer": "diffusers", "specialties": [], "training_type": "inference", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Good bug report with minimal reproducer and references to similar fixes in diffusers, but model is just an example for triggering the issue.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high" } }, { "number": 2075, "title": "lora_r is double when converting olora to lora.", "state": "closed", "author": "JaheimLee", "created_at": "2024-09-18T09:25:14Z", "updated_at": "2025-12-15T10:17:26Z", "html_url": "https://github.com/huggingface/peft/issues/2075", "labels": [], "model": "OPT", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Includes runnable reproduction with specific model, environment details, and clear expected vs actual behavior comparison.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2832, "title": "Gradient checkpoint with multiple adapters", "state": "closed", "author": "NguyenRichard", "created_at": "2025-10-14T03:53:10Z", "updated_at": "2025-12-15T08:24:03Z", "html_url": "https://github.com/huggingface/peft/issues/2832", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Includes code snippet demonstrating the adapter switching pattern and clear technical question, but no minimal reproduction or environment details.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "medium", "peft_method": "high" } }, { "number": 780, "title": "Add new method: GLoRA", "state": "closed", "author": "HsunGong", "created_at": "2023-08-03T05:37:42Z", "updated_at": "2025-12-14T15:04:19Z", "html_url": "https://github.com/huggingface/peft/issues/780", "labels": [ "contributions-welcome" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "GLoRA", "experience_score": 3, "experience_rationale": "Extremely brief feature request with only a repository link and no details.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high" } }, { "number": 2293, "title": "Is it possible to add LoRA on specific head?", "state": "closed", "author": "yuang-lee", "created_at": "2024-12-22T19:57:54Z", "updated_at": "2025-12-14T10:07:49Z", "html_url": "https://github.com/huggingface/peft/issues/2293", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 3, "experience_rationale": "minimal detail", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2892, "title": "Using trl SFTTrainer creates empty adapter.safetensors file while saving when training with LoRA and Deepspeed Zero3", "state": "closed", "author": "maximus-21", "created_at": "2025-11-04T08:09:23Z", "updated_at": "2025-12-13T15:03:53Z", "html_url": "https://github.com/huggingface/peft/issues/2892", "labels": [], "model": "unknown", "trainer": "trl", "specialties": [ "flash_attention", "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "medium", "training_type": "high", "peft_method": "high" } }, { "number": 2946, "title": "KeyError: 'token_adapter.trainable_tokens_delta.default' when saving checkpoint with trainable_token_indices and DeepSpeed ZeRO-3", "state": "closed", "author": "p1kachu2233", "created_at": "2025-12-09T16:21:46Z", "updated_at": "2025-12-12T16:08:41Z", "html_url": "https://github.com/huggingface/peft/issues/2946", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "TrainableTokens", "experience_score": 6, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "high", "peft_method": "high" } }, { "number": 2886, "title": "[FEAT] Integrate QeRL (NVFP4 + LoRA + Noise scheduler) into PEFT", "state": "closed", "author": "Aaronhuang-778", "created_at": "2025-11-03T11:42:37Z", "updated_at": "2025-12-12T15:04:57Z", "html_url": "https://github.com/huggingface/peft/issues/2886", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [ "NVFP4" ], "training_type": "RL", "peft_method": "QeRL", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "medium", "training_type": "medium", "peft_method": "high" } }, { "number": 2884, "title": "[Question/Bug] How to safely continue LoRA fine-tuning under DeepSpeed ZeRO-3 (multi-stage training with modules_to_save)", "state": "closed", "author": "XiangZhang-zx", "created_at": "2025-10-31T20:13:12Z", "updated_at": "2025-12-09T15:05:26Z", "html_url": "https://github.com/huggingface/peft/issues/2884", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "medium", "peft_method": "high" } }, { "number": 2885, "title": "PeftModel.from_pretrained and merge_and_unload don't work good for me", "state": "closed", "author": "shahelaojieraozhi", "created_at": "2025-11-01T03:17:54Z", "updated_at": "2025-12-09T15:05:24Z", "html_url": "https://github.com/huggingface/peft/issues/2885", "labels": [], "model": "Vicuna", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 5, "experience_rationale": "substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "medium" } }, { "number": 2880, "title": "Bug in fine-tuning for Qwen2.5-VL with LoRA", "state": "closed", "author": "Bytes-Lin", "created_at": "2025-10-30T07:12:30Z", "updated_at": "2025-12-08T15:05:11Z", "html_url": "https://github.com/huggingface/peft/issues/2880", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 1040, "title": "Multiple Adapters merging question", "state": "closed", "author": "AlbertoZerbinati", "created_at": "2023-10-20T10:02:31Z", "updated_at": "2025-12-08T10:59:19Z", "html_url": "https://github.com/huggingface/peft/issues/1040", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2878, "title": "peft \" target_modules='all-linear' \" have different behavior between x86 and aarch ?", "state": "closed", "author": "HuangChiEn", "created_at": "2025-10-29T03:43:02Z", "updated_at": "2025-12-07T15:03:33Z", "html_url": "https://github.com/huggingface/peft/issues/2878", "labels": [], "model": "InternVL", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "medium" } }, { "number": 2873, "title": "Can I use Lora fine-tuning twice?", "state": "closed", "author": "tohokulgq", "created_at": "2025-10-27T12:51:45Z", "updated_at": "2025-12-05T15:05:00Z", "html_url": "https://github.com/huggingface/peft/issues/2873", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2899, "title": "Prompt tuning for VLMs like Qwen2.5VL", "state": "closed", "author": "Martin9797", "created_at": "2025-11-05T16:12:11Z", "updated_at": "2025-12-04T14:19:14Z", "html_url": "https://github.com/huggingface/peft/issues/2899", "labels": [], "model": "Qwen", "trainer": "trl", "specialties": [], "training_type": "supervised", "peft_method": "Prompt Tuning", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "medium", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high" } }, { "number": 2855, "title": "When using the ZeRO3 configuration of deepspeed, the target_parameters cannot obtain the shape of the parameters", "state": "closed", "author": "qiu-xiao-0330", "created_at": "2025-10-21T04:01:45Z", "updated_at": "2025-12-01T15:05:10Z", "html_url": "https://github.com/huggingface/peft/issues/2855", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "medium", "peft_method": "medium" } }, { "number": 2834, "title": "FSDP QLORA doesn't work with multiple adapters", "state": "closed", "author": "ojh31", "created_at": "2025-10-14T14:24:44Z", "updated_at": "2025-11-30T15:03:47Z", "html_url": "https://github.com/huggingface/peft/issues/2834", "labels": [], "model": "unknown", "trainer": "trl", "specialties": [ "4bit_quantization", "bitsandbytes", "flash_attention", "FSDP" ], "training_type": "RL", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "medium", "training_type": "high", "peft_method": "high" } }, { "number": 2344, "title": "FSDP2 and peft", "state": "closed", "author": "psinger", "created_at": "2025-01-23T16:20:47Z", "updated_at": "2025-11-26T02:19:52Z", "html_url": "https://github.com/huggingface/peft/issues/2344", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "FSDP" ], "training_type": "unsure", "peft_method": "unknown", "experience_score": 4, "experience_rationale": "includes error traceback", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "low", "peft_method": "low" } }, { "number": 2914, "title": "Loading AWQ weight PeftMixedModel crashes", "state": "closed", "author": "Tfloow", "created_at": "2025-11-18T12:27:29Z", "updated_at": "2025-11-20T15:21:10Z", "html_url": "https://github.com/huggingface/peft/issues/2914", "labels": [], "model": "Llama", "trainer": "unsloth", "specialties": [ "torchao", "AWQ", "bitsandbytes", "xformers" ], "training_type": "supervised", "peft_method": "X-LoRA", "experience_score": 8, "experience_rationale": "includes code snippet; includes error traceback", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "medium", "training_type": "medium", "peft_method": "medium" } }, { "number": 2796, "title": "Support negative weights when merging LoRA adapters", "state": "closed", "author": "valteu", "created_at": "2025-09-22T13:08:40Z", "updated_at": "2025-11-19T10:50:52Z", "html_url": "https://github.com/huggingface/peft/issues/2796", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 1655, "title": "Learning the combination weights of pre-trained LoRA Modules", "state": "closed", "author": "mahdibeit", "created_at": "2024-04-15T18:25:21Z", "updated_at": "2025-11-17T23:48:27Z", "html_url": "https://github.com/huggingface/peft/issues/1655", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2901, "title": "AttributeError: 'float' object has no attribute 'meta'", "state": "closed", "author": "Karol-G", "created_at": "2025-11-06T11:24:18Z", "updated_at": "2025-11-17T15:34:08Z", "html_url": "https://github.com/huggingface/peft/issues/2901", "labels": [], "model": "ViT", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "includes code snippet; includes error traceback", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "medium" } }, { "number": 2881, "title": "Bug in prefix tuning of Qwen3-0.6B: group-query attention fix in #1901 still cause error", "state": "closed", "author": "papaya0481", "created_at": "2025-10-30T12:48:23Z", "updated_at": "2025-11-15T02:28:27Z", "html_url": "https://github.com/huggingface/peft/issues/2881", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "Prefix Tuning", "experience_score": 8, "experience_rationale": "includes code snippet; includes error traceback", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2431, "title": "Feb 2026: Update MacOS GitHub runners or remove them completely.", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-03-14T13:26:44Z", "updated_at": "2025-11-13T15:46:45Z", "html_url": "https://github.com/huggingface/peft/issues/2431", "labels": [ "wip" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 4, "experience_rationale": "includes environment details; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low" } }, { "number": 2821, "title": "PrefixTuning fails with DynamicCache: 'DynamicCache' object has no attribute 'key_cache'", "state": "closed", "author": "dnanper", "created_at": "2025-10-09T07:46:49Z", "updated_at": "2025-11-12T16:45:15Z", "html_url": "https://github.com/huggingface/peft/issues/2821", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "Prefix Tuning", "experience_score": 6, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "medium" } }, { "number": 2903, "title": "Hugging face", "state": "closed", "author": "frankjesusgonzales237-maker", "created_at": "2025-11-10T05:46:42Z", "updated_at": "2025-11-10T11:59:51Z", "html_url": "https://github.com/huggingface/peft/issues/2903", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 1, "experience_rationale": "minimal detail", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low" } }, { "number": 2699, "title": "UserWarning: Found missing adapter keys while loading the checkpoint", "state": "closed", "author": "manitadayon", "created_at": "2025-08-02T20:49:31Z", "updated_at": "2025-11-09T15:03:46Z", "html_url": "https://github.com/huggingface/peft/issues/2699", "labels": [], "model": "Llama", "trainer": "unknown", "specialties": [ "flash_attention" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "medium", "training_type": "medium", "peft_method": "high" } }, { "number": 2888, "title": "Potential remote code execution via untrusted tokenizer_kwargs in PromptEmbedding", "state": "closed", "author": "Vancir", "created_at": "2025-11-03T16:04:52Z", "updated_at": "2025-11-04T17:50:28Z", "html_url": "https://github.com/huggingface/peft/issues/2888", "labels": [], "model": "T5", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "Prompt Tuning", "experience_score": 5, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "medium" } }, { "number": 2826, "title": "Gradient checkpointing and Activated LoRA models", "state": "closed", "author": "kgreenewald", "created_at": "2025-10-09T21:24:35Z", "updated_at": "2025-11-03T12:04:41Z", "html_url": "https://github.com/huggingface/peft/issues/2826", "labels": [], "model": "unknown", "trainer": "trl", "specialties": [ "DeepSpeed" ], "training_type": "RL", "peft_method": "Activated LoRA", "experience_score": 5, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "medium", "training_type": "high", "peft_method": "high" } }, { "number": 2658, "title": "check_target_module_exists doesn't accept modules with less than 2 nesting", "state": "closed", "author": "ShirAmir", "created_at": "2025-07-20T20:04:13Z", "updated_at": "2025-10-30T15:04:36Z", "html_url": "https://github.com/huggingface/peft/issues/2658", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "medium" } }, { "number": 2795, "title": "Add Support for LoRA-XS", "state": "closed", "author": "RonanKMcGovern", "created_at": "2025-09-22T08:01:40Z", "updated_at": "2025-10-30T15:04:27Z", "html_url": "https://github.com/huggingface/peft/issues/2795", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA-XS", "experience_score": 4, "experience_rationale": "moderate detail level", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2777, "title": "Update the model config to tie_word_embeddings = False in case of lm_head or embedding layer update", "state": "closed", "author": "romitjain", "created_at": "2025-09-08T04:58:37Z", "updated_at": "2025-10-29T06:08:13Z", "html_url": "https://github.com/huggingface/peft/issues/2777", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "medium" } }, { "number": 2877, "title": "peft config 'all-linear' include lm_head, is there anyway to remove it ?", "state": "closed", "author": "HuangChiEn", "created_at": "2025-10-29T02:19:21Z", "updated_at": "2025-10-29T03:43:20Z", "html_url": "https://github.com/huggingface/peft/issues/2877", "labels": [], "model": "InternVL", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "medium" } }, { "number": 2760, "title": "Integrate DeLoRA", "state": "closed", "author": "mwbini", "created_at": "2025-08-28T23:13:00Z", "updated_at": "2025-10-28T14:35:46Z", "html_url": "https://github.com/huggingface/peft/issues/2760", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "DeLoRA", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2874, "title": "AttributeError: 'NoneType' object has no attribute 'shape'", "state": "closed", "author": "tanxiaoke6", "created_at": "2025-10-27T16:59:26Z", "updated_at": "2025-10-27T17:46:31Z", "html_url": "https://github.com/huggingface/peft/issues/2874", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 5, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low" } }, { "number": 2789, "title": "Tests fail to run: https://github.com/huggingface/peft", "state": "closed", "author": "yurivict", "created_at": "2025-09-18T20:27:05Z", "updated_at": "2025-10-27T15:04:59Z", "html_url": "https://github.com/huggingface/peft/issues/2789", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "inference", "peft_method": "unknown", "experience_score": 8, "experience_rationale": "includes error traceback; includes environment details", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "low" } }, { "number": 2787, "title": "PeftModel.from_pretrained adapter not applied when using device=\"cuda:1\"", "state": "closed", "author": "Xxxxsir", "created_at": "2025-09-18T09:04:30Z", "updated_at": "2025-10-26T15:03:46Z", "html_url": "https://github.com/huggingface/peft/issues/2787", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2833, "title": "QLORA FSDP doesn't work with 8bit", "state": "closed", "author": "ojh31", "created_at": "2025-10-14T14:06:18Z", "updated_at": "2025-10-21T13:47:53Z", "html_url": "https://github.com/huggingface/peft/issues/2833", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "8bit_quantization", "4bit_quantization", "flash_attention", "FSDP" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "high", "peft_method": "high" } }, { "number": 2636, "title": "Support for GraLoRA", "state": "closed", "author": "yeonjoon-jung01", "created_at": "2025-07-09T07:35:38Z", "updated_at": "2025-10-20T01:26:18Z", "html_url": "https://github.com/huggingface/peft/issues/2636", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "GraLoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2776, "title": "maximum recursion depth exceeded", "state": "closed", "author": "misonsky", "created_at": "2025-09-08T02:12:04Z", "updated_at": "2025-10-16T15:04:11Z", "html_url": "https://github.com/huggingface/peft/issues/2776", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 6, "experience_rationale": "includes error traceback", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low" } }, { "number": 2835, "title": "Update deprecated torch_dtype argument", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-10-14T14:50:28Z", "updated_at": "2025-10-16T12:59:10Z", "html_url": "https://github.com/huggingface/peft/issues/2835", "labels": [ "good first issue", "contributions-welcome" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 2, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low" } }, { "number": 2838, "title": "Embedding scaling for TrainableTokensModel in DoRA", "state": "closed", "author": "sambhavnoobcoder", "created_at": "2025-10-14T21:42:49Z", "updated_at": "2025-10-15T12:32:18Z", "html_url": "https://github.com/huggingface/peft/issues/2838", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "DoRA", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2840, "title": "Default set inference_mode=True For saved checkpoint", "state": "closed", "author": "dengchengxifrank", "created_at": "2025-10-15T06:08:59Z", "updated_at": "2025-10-15T12:19:42Z", "html_url": "https://github.com/huggingface/peft/issues/2840", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "bitsandbytes" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "medium", "peft_method": "medium" } }, { "number": 2802, "title": "Guide on training that requires both LoRA and base model forward calls ?", "state": "closed", "author": "thangld201", "created_at": "2025-09-27T23:12:23Z", "updated_at": "2025-10-15T10:26:15Z", "html_url": "https://github.com/huggingface/peft/issues/2802", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 527, "title": "mlp.c_proj.lora_B.default.weight has been marked as ready twice", "state": "closed", "author": "DogeWatch", "created_at": "2023-05-31T04:52:42Z", "updated_at": "2025-10-14T19:40:32Z", "html_url": "https://github.com/huggingface/peft/issues/527", "labels": [], "model": "other", "trainer": "unknown", "specialties": [ "8bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "includes code snippet; includes error traceback", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "medium", "training_type": "low", "peft_method": "high" } }, { "number": 2534, "title": "Integrate HiRA (Hadamard High-Rank Adaptation)", "state": "closed", "author": "hqsiswiliam", "created_at": "2025-05-07T15:32:31Z", "updated_at": "2025-10-14T15:04:31Z", "html_url": "https://github.com/huggingface/peft/issues/2534", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "HiRA", "experience_score": 5, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2830, "title": "Embedding scaling for TrainableTokensModel in XLoraEmbeddingLayer", "state": "closed", "author": "sambhavnoobcoder", "created_at": "2025-10-13T06:31:55Z", "updated_at": "2025-10-14T14:06:12Z", "html_url": "https://github.com/huggingface/peft/issues/2830", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "X-LoRA", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2809, "title": "TrainableTokensModel doesn't handle embedding scaling", "state": "closed", "author": "kimihailv", "created_at": "2025-10-01T18:11:33Z", "updated_at": "2025-10-14T10:35:32Z", "html_url": "https://github.com/huggingface/peft/issues/2809", "labels": [], "model": "Gemma", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "TrainableTokens", "experience_score": 5, "experience_rationale": "substantial description", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2745, "title": "Support QLoRA Training on AWQ Quantized Models", "state": "closed", "author": "yhhtc1201", "created_at": "2025-08-19T05:19:26Z", "updated_at": "2025-10-13T15:04:48Z", "html_url": "https://github.com/huggingface/peft/issues/2745", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "AWQ", "bitsandbytes" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "moderate detail level", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "medium", "peft_method": "high" } }, { "number": 2769, "title": "prepare_inputs_for_generation missing when using newer transformers version", "state": "closed", "author": "Hrovatin", "created_at": "2025-09-04T05:56:24Z", "updated_at": "2025-10-13T15:04:46Z", "html_url": "https://github.com/huggingface/peft/issues/2769", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes environment details; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "medium" } }, { "number": 2766, "title": "[X-LoRA] RuntimeError: Tensor Size Mismatch with Vision-Language Models due to Sequence Length Inference", "state": "closed", "author": "seongbin1231", "created_at": "2025-09-03T12:38:34Z", "updated_at": "2025-10-12T15:03:22Z", "html_url": "https://github.com/huggingface/peft/issues/2766", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "X-LoRA", "experience_score": 6, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2827, "title": "ImportError: `PytorchGELUTanh` Removed in Transformers 4.57.0 but Still Required by peft", "state": "closed", "author": "Sunhill666", "created_at": "2025-10-10T05:53:58Z", "updated_at": "2025-10-10T09:01:47Z", "html_url": "https://github.com/huggingface/peft/issues/2827", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [ "AWQ" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "includes code snippet; includes error traceback", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "medium", "training_type": "low", "peft_method": "medium" } }, { "number": 2764, "title": "merge_and_unload returns the base (prior to fine-tuning) back!!!!", "state": "closed", "author": "manitadayon", "created_at": "2025-09-01T04:07:36Z", "updated_at": "2025-10-09T15:26:15Z", "html_url": "https://github.com/huggingface/peft/issues/2764", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "unknown", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "low" } }, { "number": 2786, "title": "Two issues with [X-LoRA] Implementation: Scalings Logging and Top-K Softmax Normalization", "state": "closed", "author": "Che-Xu", "created_at": "2025-09-17T09:01:29Z", "updated_at": "2025-10-09T11:36:56Z", "html_url": "https://github.com/huggingface/peft/issues/2786", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [ "DeepSpeed" ], "training_type": "unsure", "peft_method": "X-LoRA", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "medium", "training_type": "low", "peft_method": "high" } }, { "number": 2816, "title": "PromptEmbedding's random initialization leads to out-of-manifold embeddings that doesn't learn", "state": "closed", "author": "macmacmacmac", "created_at": "2025-10-07T18:50:34Z", "updated_at": "2025-10-09T10:25:21Z", "html_url": "https://github.com/huggingface/peft/issues/2816", "labels": [], "model": "Llama", "trainer": "unknown", "specialties": [], "training_type": "inference", "peft_method": "unknown", "experience_score": 2, "experience_rationale": "substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "low" } }, { "number": 2822, "title": "Qwen2.5 VL加载checkpoint继续训练报错", "state": "closed", "author": "Yimi81", "created_at": "2025-10-09T09:13:21Z", "updated_at": "2025-10-09T09:25:24Z", "html_url": "https://github.com/huggingface/peft/issues/2822", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [ "flash_attention", "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "medium", "training_type": "high", "peft_method": "medium" } }, { "number": 2552, "title": "WaveFT Integration", "state": "closed", "author": "Bilican", "created_at": "2025-05-21T15:38:16Z", "updated_at": "2025-10-09T04:42:48Z", "html_url": "https://github.com/huggingface/peft/issues/2552", "labels": [ "wip" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "AdaLoRA", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2179, "title": "Integration of merge-kit into PEFT ", "state": "closed", "author": "ParagEkbote", "created_at": "2024-10-25T19:53:14Z", "updated_at": "2025-10-07T09:30:41Z", "html_url": "https://github.com/huggingface/peft/issues/2179", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "unknown", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "low" } }, { "number": 2566, "title": "Found missing adapter keys while loading the checkpoint", "state": "closed", "author": "JJplane", "created_at": "2025-06-03T09:56:08Z", "updated_at": "2025-10-06T20:32:14Z", "html_url": "https://github.com/huggingface/peft/issues/2566", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2759, "title": "PeftModel trainable parameters with multiple adapters", "state": "closed", "author": "NguyenRichard", "created_at": "2025-08-28T16:36:25Z", "updated_at": "2025-10-06T15:04:09Z", "html_url": "https://github.com/huggingface/peft/issues/2759", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 5, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 1142, "title": "How to do Gradient Checkpoint + LoRA", "state": "closed", "author": "tcapelle", "created_at": "2023-11-17T09:34:16Z", "updated_at": "2025-10-06T10:22:58Z", "html_url": "https://github.com/huggingface/peft/issues/1142", "labels": [], "model": "Llama", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high" } }, { "number": 2516, "title": "Call for contribution: KaSA: Knowledge-Aware Singular-Value Adaptation of Large Language Models", "state": "open", "author": "BenjaminBossan", "created_at": "2025-04-25T13:59:34Z", "updated_at": "2025-10-06T09:49:17Z", "html_url": "https://github.com/huggingface/peft/issues/2516", "labels": [ "wip" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "DoRA", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2772, "title": "\"UserWarning: Found missing adapter keys while loading the checkpoint\" when module name contains \"weight\"", "state": "closed", "author": "mauricebrenner", "created_at": "2025-09-05T11:43:29Z", "updated_at": "2025-09-30T09:09:46Z", "html_url": "https://github.com/huggingface/peft/issues/2772", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2749, "title": "Set multiple adapters actively when training", "state": "closed", "author": "liaoyy-hub", "created_at": "2025-08-21T09:59:25Z", "updated_at": "2025-09-29T15:04:15Z", "html_url": "https://github.com/huggingface/peft/issues/2749", "labels": [], "model": "unknown", "trainer": "transformers", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high" } }, { "number": 2757, "title": "Remove usage of from_legacy_cache", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-08-28T12:39:31Z", "updated_at": "2025-09-29T09:29:53Z", "html_url": "https://github.com/huggingface/peft/issues/2757", "labels": [ "contributions-welcome" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 3, "experience_rationale": "includes environment details", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low" } }, { "number": 2783, "title": "PEFT Loha/Lokr crashes for SDXL pipeline inference", "state": "closed", "author": "frutiemax92", "created_at": "2025-09-13T21:12:37Z", "updated_at": "2025-09-25T15:59:46Z", "html_url": "https://github.com/huggingface/peft/issues/2783", "labels": [], "model": "Stable Diffusion", "trainer": "unknown", "specialties": [], "training_type": "inference", "peft_method": "LoHa", "experience_score": 7, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2742, "title": "RuntimeError: element 0 of tensors does not require grad and does not have a grad_fn", "state": "closed", "author": "Mishajain1110", "created_at": "2025-08-15T06:21:50Z", "updated_at": "2025-09-23T15:04:07Z", "html_url": "https://github.com/huggingface/peft/issues/2742", "labels": [], "model": "Llama", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "unknown", "experience_score": 6, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "low" } }, { "number": 2798, "title": "Add stricter type checking in LoraConfig for support with HfArgumentParser", "state": "closed", "author": "romitjain", "created_at": "2025-09-23T05:19:34Z", "updated_at": "2025-09-23T12:37:47Z", "html_url": "https://github.com/huggingface/peft/issues/2798", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "DeepSpeed" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 5, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "low", "peft_method": "high" } }, { "number": 2653, "title": "trainable_token_indices causes training to fail", "state": "closed", "author": "mark-nlp", "created_at": "2025-07-17T20:56:48Z", "updated_at": "2025-09-22T15:04:45Z", "html_url": "https://github.com/huggingface/peft/issues/2653", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "TrainableTokens", "experience_score": 7, "experience_rationale": "includes code snippet; includes error traceback", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "medium" } }, { "number": 2673, "title": "ModuleNotFoundError: No module named 'peft'", "state": "closed", "author": "SAOHEIUD", "created_at": "2025-07-27T23:23:32Z", "updated_at": "2025-09-19T15:04:08Z", "html_url": "https://github.com/huggingface/peft/issues/2673", "labels": [], "model": "Mistral", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 6, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low" } }, { "number": 2523, "title": "Support for Activated-Lora", "state": "closed", "author": "vpgits", "created_at": "2025-04-29T10:23:19Z", "updated_at": "2025-09-18T13:38:58Z", "html_url": "https://github.com/huggingface/peft/issues/2523", "labels": [ "wip" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "inference", "peft_method": "Activated LoRA", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2733, "title": "LoraLayer.set_scale not working correctly with rslora", "state": "closed", "author": "dongmingli-Ben", "created_at": "2025-08-08T19:08:49Z", "updated_at": "2025-09-16T09:30:58Z", "html_url": "https://github.com/huggingface/peft/issues/2733", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 1593, "title": "Getting Dora Model Is Very Slow", "state": "closed", "author": "mallorbc", "created_at": "2024-03-27T06:05:56Z", "updated_at": "2025-09-10T23:41:37Z", "html_url": "https://github.com/huggingface/peft/issues/1593", "labels": [], "model": "Mistral", "trainer": "unknown", "specialties": [ "flash_attention", "bitsandbytes", "DeepSpeed" ], "training_type": "unsure", "peft_method": "DoRA", "experience_score": 6, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "medium", "training_type": "low", "peft_method": "high" } }, { "number": 2683, "title": "Add Fast-Forward to PEFT", "state": "closed", "author": "AdirRahamim", "created_at": "2025-07-30T20:14:03Z", "updated_at": "2025-09-08T15:03:47Z", "html_url": "https://github.com/huggingface/peft/issues/2683", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "DoRA", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high" } }, { "number": 1358, "title": "AttributeError: 'LoraModel' object has no attribute 'prepare_inputs_for_generation'", "state": "closed", "author": "belegubesa", "created_at": "2024-01-13T14:31:18Z", "updated_at": "2025-09-06T17:16:55Z", "html_url": "https://github.com/huggingface/peft/issues/1358", "labels": [], "model": "Llama", "trainer": "trl", "specialties": [ "bitsandbytes" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "includes environment details; substantial description", "confidence": { "model": "medium", "trainer_framework": "high", "specialties": "medium", "training_type": "high", "peft_method": "high" } }, { "number": 2652, "title": "Add SingLoRA to PEFT", "state": "closed", "author": "DavidBCode", "created_at": "2025-07-17T17:39:20Z", "updated_at": "2025-09-03T04:13:14Z", "html_url": "https://github.com/huggingface/peft/issues/2652", "labels": [], "model": "Llama", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "SingLoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 1472, "title": "Integrating X-LoRA", "state": "closed", "author": "EricLBuehler", "created_at": "2024-02-16T13:13:17Z", "updated_at": "2025-09-03T04:00:50Z", "html_url": "https://github.com/huggingface/peft/issues/1472", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "X-LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2688, "title": "[DOC] OFT and BOFT descriptions are the same", "state": "closed", "author": "AmirBachir", "created_at": "2025-07-31T23:04:20Z", "updated_at": "2025-09-02T09:32:26Z", "html_url": "https://github.com/huggingface/peft/issues/2688", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "OFT", "experience_score": 3, "experience_rationale": "minimal detail", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2755, "title": "EfficientQAT", "state": "closed", "author": "gapsong", "created_at": "2025-08-27T17:07:22Z", "updated_at": "2025-09-02T09:24:38Z", "html_url": "https://github.com/huggingface/peft/issues/2755", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "EfficientQAT", "experience_score": 1, "experience_rationale": "minimal detail", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 876, "title": "Performance of Reloaded Models are Much Worse than the Fine-Tuned Model", "state": "closed", "author": "houx15", "created_at": "2023-08-29T09:19:08Z", "updated_at": "2025-09-01T09:30:10Z", "html_url": "https://github.com/huggingface/peft/issues/876", "labels": [], "model": "Llama", "trainer": "transformers", "specialties": [ "8bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes code snippet; substantial description", "confidence": { "model": "medium", "trainer_framework": "high", "specialties": "medium", "training_type": "high", "peft_method": "medium" } }, { "number": 1929, "title": "Pissa's [path_initial_model_for_weight_conversion] to build LoRA on top of base model", "state": "closed", "author": "mtran14", "created_at": "2024-07-16T10:51:53Z", "updated_at": "2025-08-29T02:31:14Z", "html_url": "https://github.com/huggingface/peft/issues/1929", "labels": [], "model": "GPT", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2756, "title": "Loading multiple adapters with PeftMixedModel", "state": "closed", "author": "NguyenRichard", "created_at": "2025-08-27T22:36:53Z", "updated_at": "2025-08-28T14:56:29Z", "html_url": "https://github.com/huggingface/peft/issues/2756", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "high" } }, { "number": 2754, "title": "No working peft version available in PyPI for transformers 4.55+ version", "state": "closed", "author": "qGentry", "created_at": "2025-08-27T16:19:16Z", "updated_at": "2025-08-28T13:33:05Z", "html_url": "https://github.com/huggingface/peft/issues/2754", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "vLLM" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "medium", "peft_method": "medium" } }, { "number": 2650, "title": "AttributeError: 'DTensor' object has no attribute 'compress_statistics'", "state": "closed", "author": "vsoesanto", "created_at": "2025-07-16T01:02:07Z", "updated_at": "2025-08-27T15:04:40Z", "html_url": "https://github.com/huggingface/peft/issues/2650", "labels": [], "model": "unknown", "trainer": "trl", "specialties": [ "4bit_quantization", "8bit_quantization", "bitsandbytes" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "medium", "training_type": "high", "peft_method": "medium" } }, { "number": 1935, "title": "[Call for contributions] help us improve LoKr, LoHa, and other LyCORIS ", "state": "closed", "author": "sayakpaul", "created_at": "2024-07-18T04:42:18Z", "updated_at": "2025-08-27T11:07:07Z", "html_url": "https://github.com/huggingface/peft/issues/1935", "labels": [ "wip", "contributions-welcome", "good-second-pr" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoHa", "experience_score": 3, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2746, "title": "Gemma 2/3 Attention: Expected a single attention mask, got 2 instead", "state": "closed", "author": "michelleezhang", "created_at": "2025-08-20T18:08:02Z", "updated_at": "2025-08-27T02:43:22Z", "html_url": "https://github.com/huggingface/peft/issues/2746", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "Prompt Tuning", "experience_score": 5, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "medium" } }, { "number": 2751, "title": "Number of trainable params is displayed wrong in accelerate logs when using FSDP", "state": "closed", "author": "Ali-Sayed-Salehi", "created_at": "2025-08-25T22:44:15Z", "updated_at": "2025-08-26T16:29:01Z", "html_url": "https://github.com/huggingface/peft/issues/2751", "labels": [], "model": "Llama", "trainer": "trl", "specialties": [ "4bit_quantization", "8bit_quantization", "DeepSpeed", "bitsandbytes", "FSDP" ], "training_type": "supervised", "peft_method": "unknown", "experience_score": 8, "experience_rationale": "includes code snippet; includes error traceback", "confidence": { "model": "medium", "trainer_framework": "high", "specialties": "medium", "training_type": "medium", "peft_method": "low" } }, { "number": 2654, "title": "Implementation of `Bottleneck`, or known as `Adapter`", "state": "closed", "author": "Zuo-Lihan", "created_at": "2025-07-18T01:07:38Z", "updated_at": "2025-08-25T15:04:34Z", "html_url": "https://github.com/huggingface/peft/issues/2654", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "4bit_quantization", "8bit_quantization", "DeepSpeed", "bitsandbytes", "FSDP" ], "training_type": "supervised", "peft_method": "Bottleneck", "experience_score": 7, "experience_rationale": "includes error traceback; substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "high", "peft_method": "high" } }, { "number": 2655, "title": "Request: Enable LoRA support for additional/custom layers beyond Linear, Conv, and Embedding", "state": "closed", "author": "asrays", "created_at": "2025-07-18T05:43:13Z", "updated_at": "2025-08-25T15:04:31Z", "html_url": "https://github.com/huggingface/peft/issues/2655", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 3, "experience_rationale": "minimal detail", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high" } }, { "number": 2672, "title": "Support for RoAD: 2D Rotary Adaptation for Efficient Finetuning, Efficient Batching and Composability", "state": "closed", "author": "ppetrushkov", "created_at": "2025-07-25T12:01:13Z", "updated_at": "2025-08-25T09:31:47Z", "html_url": "https://github.com/huggingface/peft/issues/2672", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "vLLM" ], "training_type": "supervised", "peft_method": "OFT", "experience_score": 4, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "medium", "training_type": "medium", "peft_method": "high" } }, { "number": 2642, "title": "Prompt_Tuning.ipynb example doesn't seem to train the model", "state": "closed", "author": "ruixing76", "created_at": "2025-07-11T18:26:58Z", "updated_at": "2025-08-23T15:03:47Z", "html_url": "https://github.com/huggingface/peft/issues/2642", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "Prompt Tuning", "experience_score": 2, "experience_rationale": "substantial description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "medium" } }, { "number": 2647, "title": "How can I merge the original model weights with LoRA weights?", "state": "closed", "author": "guoguo1314", "created_at": "2025-07-15T11:40:33Z", "updated_at": "2025-08-23T15:03:44Z", "html_url": "https://github.com/huggingface/peft/issues/2647", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2747, "title": "Add support for DINO v3", "state": "closed", "author": "Six-Bit-TX", "created_at": "2025-08-21T07:52:18Z", "updated_at": "2025-08-21T09:35:05Z", "html_url": "https://github.com/huggingface/peft/issues/2747", "labels": [], "model": "DINO", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2695, "title": "TST: Some SHiRA tests are failing when run with parallelism", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-08-01T17:16:04Z", "updated_at": "2025-08-19T09:55:26Z", "html_url": "https://github.com/huggingface/peft/issues/2695", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "SHiRA", "experience_score": 10, "experience_rationale": "Issue has clear title, system info provided.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2627, "title": "Some Gemma generate tests fail on GPU with: torch._dynamo.exc.Unsupported: Data-dependent branching", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-07-02T09:21:19Z", "updated_at": "2025-08-19T09:36:24Z", "html_url": "https://github.com/huggingface/peft/issues/2627", "labels": [ "bug" ], "model": "Gemma", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2640, "title": "Why peft.utils.other.fsdp_auto_wrap_policy do not warp the module do not require grad?", "state": "closed", "author": "Changlin-Lee", "created_at": "2025-07-10T12:07:13Z", "updated_at": "2025-08-18T15:05:03Z", "html_url": "https://github.com/huggingface/peft/issues/2640", "labels": [], "model": "unknown", "trainer": "none", "specialties": [ "FSDP" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2556, "title": "[LORA Mamba] Raise Error When out_proj or conv1d Are Included in LoRA Config for Mamba-Based Models (e.g. Falcon-H1)", "state": "closed", "author": "dhiaEddineRhaiem", "created_at": "2025-05-29T08:31:08Z", "updated_at": "2025-08-14T13:05:34Z", "html_url": "https://github.com/huggingface/peft/issues/2556", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium" } }, { "number": 2738, "title": "Which base model weights are getting frozen after applying LoRA?", "state": "closed", "author": "srbh-dl", "created_at": "2025-08-13T17:35:10Z", "updated_at": "2025-08-14T04:20:42Z", "html_url": "https://github.com/huggingface/peft/issues/2738", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Issue has clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 1750, "title": "How to finetune embeddings and LM head as a single layer when they are tied?", "state": "closed", "author": "GokulNC", "created_at": "2024-05-21T18:32:07Z", "updated_at": "2025-08-12T11:54:09Z", "html_url": "https://github.com/huggingface/peft/issues/1750", "labels": [], "model": "Gemma", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Issue has detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 506, "title": "ValueError: The following model_kwargs are not used by the model: ['token_type_ids'] (note: typos in the generate arguments will also show up in this list)", "state": "closed", "author": "Imran-tech-lab", "created_at": "2023-05-27T00:44:31Z", "updated_at": "2025-08-08T14:48:46Z", "html_url": "https://github.com/huggingface/peft/issues/506", "labels": [], "model": "Falcon", "trainer": "unknown", "specialties": [], "training_type": "inference", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "Issue has error details provided, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "low", "experience_score": "medium" } }, { "number": 2716, "title": "Lora_bias: NoneType object has no attribute data", "state": "closed", "author": "Fhrozen", "created_at": "2025-08-06T10:44:11Z", "updated_at": "2025-08-07T12:26:29Z", "html_url": "https://github.com/huggingface/peft/issues/2716", "labels": [], "model": "unknown", "trainer": "SFTTrainer", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium" } }, { "number": 2527, "title": "LoRA for experts layers in MoE", "state": "closed", "author": "ebsmothers", "created_at": "2025-05-01T16:53:22Z", "updated_at": "2025-08-06T09:41:54Z", "html_url": "https://github.com/huggingface/peft/issues/2527", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue has detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2690, "title": "failed to run kaggle notebook", "state": "closed", "author": "MengAiDev", "created_at": "2025-08-01T09:50:44Z", "updated_at": "2025-08-02T03:27:57Z", "html_url": "https://github.com/huggingface/peft/issues/2690", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "inference", "peft_method": "Adapters", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "medium", "peft_method": "high", "experience_score": "medium" } }, { "number": 808, "title": "What is the correct way to apply LoRA on a custom model (not models on HuggingFace)?", "state": "closed", "author": "DtYXs", "created_at": "2023-08-08T17:10:36Z", "updated_at": "2025-08-01T21:14:25Z", "html_url": "https://github.com/huggingface/peft/issues/808", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2608, "title": "support lora weights conversion", "state": "closed", "author": "garychan22", "created_at": "2025-06-24T09:27:42Z", "updated_at": "2025-08-01T15:04:47Z", "html_url": "https://github.com/huggingface/peft/issues/2608", "labels": [], "model": "Stable Diffusion", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2321, "title": "[Warning] `Merge lora module to 4-bit linear may get different generations`", "state": "closed", "author": "steveepreston", "created_at": "2025-01-11T20:27:54Z", "updated_at": "2025-07-29T16:09:05Z", "html_url": "https://github.com/huggingface/peft/issues/2321", "labels": [], "model": "Gemma", "trainer": "peft_direct", "specialties": [ "bitsandbytes", "4bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 1569, "title": "Error saving LoRaModel with Wav2vec2forCTC basemodel", "state": "closed", "author": "geoffvdr", "created_at": "2024-03-18T22:37:58Z", "updated_at": "2025-07-29T15:06:27Z", "html_url": "https://github.com/huggingface/peft/issues/1569", "labels": [], "model": "Wav2Vec2/Whisper", "trainer": "transformers.Trainer", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 1720, "title": "RuntimeError: only Tensors of floating point dtype can require gradients for QLoRA since transformers 4.40", "state": "closed", "author": "dipanjanS", "created_at": "2024-05-09T14:51:11Z", "updated_at": "2025-07-29T06:58:44Z", "html_url": "https://github.com/huggingface/peft/issues/1720", "labels": [], "model": "BERT", "trainer": "unknown", "specialties": [ "bitsandbytes", "4bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2595, "title": "LoRA combined with Prefix tuning", "state": "closed", "author": "aparaselli", "created_at": "2025-06-18T14:42:23Z", "updated_at": "2025-07-28T15:04:48Z", "html_url": "https://github.com/huggingface/peft/issues/2595", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "Prefix Tuning", "experience_score": 7, "experience_rationale": "Issue has clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 1672, "title": "Using PEFT causes model to not predict EOS", "state": "closed", "author": "Km3888", "created_at": "2024-04-23T18:23:50Z", "updated_at": "2025-07-27T12:45:42Z", "html_url": "https://github.com/huggingface/peft/issues/1672", "labels": [], "model": "GPT", "trainer": "SFTTrainer", "specialties": [ "DeepSpeed", "FSDP", "bitsandbytes", "4bit_quantization", "8bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows detailed description, clear title, system info provided.", "confidence": { "model": "medium", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2587, "title": "Support for ABBA: Highly Expressive Hadamard Product Adaptation for LLMs", "state": "closed", "author": "RaghavSinghal10", "created_at": "2025-06-12T14:35:26Z", "updated_at": "2025-07-26T15:04:07Z", "html_url": "https://github.com/huggingface/peft/issues/2587", "labels": [], "model": "Gemma", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "DoRA", "experience_score": 8, "experience_rationale": "Issue has detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2671, "title": "QLora broken - bnb.py resolve_variant should accept QALoRA params for compatibility with LoraLayer", "state": "closed", "author": "gapsong", "created_at": "2025-07-25T09:53:13Z", "updated_at": "2025-07-25T10:36:11Z", "html_url": "https://github.com/huggingface/peft/issues/2671", "labels": [], "model": "unknown", "trainer": "trl", "specialties": [ "bitsandbytes", "4bit_quantization" ], "training_type": "supervised", "peft_method": "QaLoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium" } }, { "number": 2660, "title": "Custom models LoRA", "state": "closed", "author": "stillbetter", "created_at": "2025-07-21T11:52:30Z", "updated_at": "2025-07-24T12:53:34Z", "html_url": "https://github.com/huggingface/peft/issues/2660", "labels": [], "model": "Qwen", "trainer": "llamafactory", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Issue has limited context with 234 chars body.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2434, "title": "Request for integrate sine-LoRA", "state": "open", "author": "yipingji", "created_at": "2025-03-18T15:49:34Z", "updated_at": "2025-07-24T05:40:45Z", "html_url": "https://github.com/huggingface/peft/issues/2434", "labels": [ "wip" ], "model": "Llama", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "DoRA", "experience_score": 9, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2663, "title": "[Bug] PiSSA example script incompatible with current trl dependency", "state": "closed", "author": "gapsong", "created_at": "2025-07-22T12:08:51Z", "updated_at": "2025-07-22T13:57:49Z", "html_url": "https://github.com/huggingface/peft/issues/2663", "labels": [], "model": "unknown", "trainer": "SFTTrainer", "specialties": [], "training_type": "unsure", "peft_method": "PiSSA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2613, "title": "Make Task Type validation case insensitive", "state": "closed", "author": "cartesian-plane", "created_at": "2025-06-25T14:07:36Z", "updated_at": "2025-07-22T12:07:16Z", "html_url": "https://github.com/huggingface/peft/issues/2613", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 9, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 854, "title": "Sharing backbone model for different task adapters", "state": "closed", "author": "jedzill4", "created_at": "2023-08-23T17:22:44Z", "updated_at": "2025-07-22T10:22:21Z", "html_url": "https://github.com/huggingface/peft/issues/854", "labels": [], "model": "BERT", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Issue has detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2656, "title": "Fix BibTeX", "state": "closed", "author": "cx-alberto-simoes", "created_at": "2025-07-18T07:42:17Z", "updated_at": "2025-07-21T12:36:03Z", "html_url": "https://github.com/huggingface/peft/issues/2656", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 9, "experience_rationale": "Issue shows includes code snippet, detailed description, system info provided.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2583, "title": "Lora transfer learning", "state": "closed", "author": "hardikdava", "created_at": "2025-06-11T12:00:25Z", "updated_at": "2025-07-20T15:04:05Z", "html_url": "https://github.com/huggingface/peft/issues/2583", "labels": [], "model": "FLUX", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2436, "title": "Fine-tuning with Multiple LoRAs", "state": "closed", "author": "xymou", "created_at": "2025-03-19T13:49:28Z", "updated_at": "2025-07-19T05:45:12Z", "html_url": "https://github.com/huggingface/peft/issues/2436", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2615, "title": "How can I fine-tune the linear layers of the LLM part in Qwen2.5_VL 3B?", "state": "closed", "author": "guoguo1314", "created_at": "2025-06-26T02:08:43Z", "updated_at": "2025-07-18T16:04:27Z", "html_url": "https://github.com/huggingface/peft/issues/2615", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2589, "title": "Clarification on embed tokens when using LoRA", "state": "closed", "author": "jiosephlee", "created_at": "2025-06-13T18:46:45Z", "updated_at": "2025-07-17T16:16:52Z", "html_url": "https://github.com/huggingface/peft/issues/2589", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Issue has clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2558, "title": "GraLoRA support?", "state": "closed", "author": "DiTo97", "created_at": "2025-05-29T18:36:27Z", "updated_at": "2025-07-15T15:04:20Z", "html_url": "https://github.com/huggingface/peft/issues/2558", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "GraLoRA", "experience_score": 7, "experience_rationale": "Issue has detailed description.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2585, "title": "Sparse High Rank Adapters (SHiRA) Support", "state": "closed", "author": "kkb-code", "created_at": "2025-06-11T18:26:37Z", "updated_at": "2025-07-15T09:00:56Z", "html_url": "https://github.com/huggingface/peft/issues/2585", "labels": [ "wip" ], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "SHiRA", "experience_score": 7, "experience_rationale": "Issue has detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2639, "title": "Multi-Lora training", "state": "closed", "author": "liumc14", "created_at": "2025-07-10T09:21:36Z", "updated_at": "2025-07-14T14:23:00Z", "html_url": "https://github.com/huggingface/peft/issues/2639", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue has includes code snippet, detailed description.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2545, "title": "GraniteMoE’s implementation is not compatible with HF’s peft", "state": "closed", "author": "lhkhiem28", "created_at": "2025-03-03T18:01:57Z", "updated_at": "2025-07-14T08:54:37Z", "html_url": "https://github.com/huggingface/peft/issues/2545", "labels": [], "model": "Granite", "trainer": "none", "specialties": [ "vLLM" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2634, "title": "Auto-tagging does not support customized peft type", "state": "closed", "author": "criyle", "created_at": "2025-07-06T19:09:14Z", "updated_at": "2025-07-08T09:13:50Z", "html_url": "https://github.com/huggingface/peft/issues/2634", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2557, "title": "ValueError during evaluation when using label_names=['labels']", "state": "closed", "author": "Cryptocho", "created_at": "2025-05-29T10:21:55Z", "updated_at": "2025-07-07T15:04:00Z", "html_url": "https://github.com/huggingface/peft/issues/2557", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [ "bitsandbytes", "4bit_quantization" ], "training_type": "unsure", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2449, "title": "the bug of Prefix tuning", "state": "closed", "author": "aohenuo", "created_at": "2025-03-25T05:57:25Z", "updated_at": "2025-07-07T09:18:11Z", "html_url": "https://github.com/huggingface/peft/issues/2449", "labels": [], "model": "Llama", "trainer": "transformers.Trainer", "specialties": [], "training_type": "supervised", "peft_method": "Prefix Tuning", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium" } }, { "number": 986, "title": "QaLora", "state": "closed", "author": "thistleknot", "created_at": "2023-10-01T23:48:27Z", "updated_at": "2025-07-03T16:23:56Z", "html_url": "https://github.com/huggingface/peft/issues/986", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "QaLoRA", "experience_score": 8, "experience_rationale": "Issue has limited context with 335 chars body.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2540, "title": "NotImplementedError While prefix tuning the Gemma 4b model", "state": "closed", "author": "imcoza", "created_at": "2025-05-12T09:04:11Z", "updated_at": "2025-07-02T15:04:28Z", "html_url": "https://github.com/huggingface/peft/issues/2540", "labels": [], "model": "Gemma", "trainer": "transformers.Trainer", "specialties": [ "DeepSpeed", "flash_attention" ], "training_type": "supervised", "peft_method": "Prefix Tuning", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium" } }, { "number": 2546, "title": "'PeftModelForCausalLM' object has no attribute 'save_pretrained_merged'", "state": "closed", "author": "Z1176893827", "created_at": "2025-05-19T15:15:57Z", "updated_at": "2025-07-02T15:04:26Z", "html_url": "https://github.com/huggingface/peft/issues/2546", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 6, "experience_rationale": "Issue has clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2549, "title": "DoRA RuntimeError: shape is invalid when Conv2d(groups=2)", "state": "closed", "author": "victolee0", "created_at": "2025-05-20T05:57:45Z", "updated_at": "2025-06-30T14:50:56Z", "html_url": "https://github.com/huggingface/peft/issues/2549", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "DoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2467, "title": "File update: use LoRa with PEFT and DeepSpeed documentation is already lagging behind", "state": "closed", "author": "pengjunfeng11", "created_at": "2025-03-29T08:52:28Z", "updated_at": "2025-06-29T15:04:07Z", "html_url": "https://github.com/huggingface/peft/issues/2467", "labels": [], "model": "unknown", "trainer": "none", "specialties": [ "DeepSpeed", "4bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Issue has detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2603, "title": "lora with trainable_token_indices do NOT support zero3？", "state": "closed", "author": "LinB203", "created_at": "2025-06-22T13:38:13Z", "updated_at": "2025-06-26T14:48:00Z", "html_url": "https://github.com/huggingface/peft/issues/2603", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "high", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2607, "title": "QLoRA tuning support for compressed-tensors", "state": "closed", "author": "kmehant", "created_at": "2025-06-24T08:19:07Z", "updated_at": "2025-06-26T08:47:30Z", "html_url": "https://github.com/huggingface/peft/issues/2607", "labels": [], "model": "unknown", "trainer": "none", "specialties": [ "4bit_quantization" ], "training_type": "unsure", "peft_method": "unknown", "experience_score": 8, "experience_rationale": "Issue has clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 816, "title": "Question: why does prepare_model_for_kbit_training cast input **and** output embeddings to float32?", "state": "closed", "author": "Birch-san", "created_at": "2023-08-09T16:07:03Z", "updated_at": "2025-06-25T13:30:05Z", "html_url": "https://github.com/huggingface/peft/issues/816", "labels": [], "model": "Llama", "trainer": "unknown", "specialties": [ "4bit_quantization" ], "training_type": "unsure", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2592, "title": "diffusion components have no 'add_weighted_adapter'", "state": "closed", "author": "garychan22", "created_at": "2025-06-17T10:00:09Z", "updated_at": "2025-06-23T12:53:50Z", "html_url": "https://github.com/huggingface/peft/issues/2592", "labels": [], "model": "Stable Diffusion", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2601, "title": "How to Load Adapters with Per-Layer Variable Shapes in `PeftModel.from_pretrained`", "state": "closed", "author": "yuxuan-z19", "created_at": "2025-06-20T11:11:19Z", "updated_at": "2025-06-21T05:42:58Z", "html_url": "https://github.com/huggingface/peft/issues/2601", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 7, "experience_rationale": "Issue has detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2533, "title": "Integrate TLoRA (Tri-Matrix LoRA)", "state": "closed", "author": "itanvir", "created_at": "2025-05-06T21:22:50Z", "updated_at": "2025-06-15T15:03:57Z", "html_url": "https://github.com/huggingface/peft/issues/2533", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "TLoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 1612, "title": "LoRA layer param.grad=None after loss.backward() using lora for mamba 2.8B", "state": "closed", "author": "jinlHe", "created_at": "2024-04-01T15:06:12Z", "updated_at": "2025-06-12T10:34:25Z", "html_url": "https://github.com/huggingface/peft/issues/1612", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2586, "title": "Merging QLoRA weights to Qwen-2.5-VL-7B-Instruct does not change base model weights", "state": "closed", "author": "harshit2997", "created_at": "2025-06-12T00:50:23Z", "updated_at": "2025-06-12T03:15:49Z", "html_url": "https://github.com/huggingface/peft/issues/2586", "labels": [], "model": "Qwen", "trainer": "SFTTrainer", "specialties": [ "bitsandbytes", "flash_attention", "4bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2518, "title": "Multi-gpu inference llama-3.2 vision with QLoRA", "state": "closed", "author": "shon-otmazgin-wix", "created_at": "2025-04-27T16:16:06Z", "updated_at": "2025-06-05T15:04:34Z", "html_url": "https://github.com/huggingface/peft/issues/2518", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [ "bitsandbytes", "4bit_quantization" ], "training_type": "inference", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "high", "training_type": "medium", "peft_method": "low", "experience_score": "medium" } }, { "number": 2541, "title": "Inconsistent variable naming makes bugs harder to understand", "state": "closed", "author": "pranav-gade", "created_at": "2025-05-15T10:16:36Z", "updated_at": "2025-06-04T15:37:54Z", "html_url": "https://github.com/huggingface/peft/issues/2541", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 376, "title": "FineTuning BLIP2 - various issues", "state": "closed", "author": "ziggygram", "created_at": "2023-04-27T21:58:57Z", "updated_at": "2025-06-02T12:32:50Z", "html_url": "https://github.com/huggingface/peft/issues/376", "labels": [], "model": "T5", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium" } }, { "number": 2538, "title": "isthere a way to merge model after use P-Tuning", "state": "closed", "author": "lixd", "created_at": "2025-05-09T09:02:36Z", "updated_at": "2025-05-30T01:14:19Z", "html_url": "https://github.com/huggingface/peft/issues/2538", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2501, "title": "merge_adapter does not work with bitsandbytes 4bit quantized model.", "state": "closed", "author": "lfranceschetti", "created_at": "2025-04-16T11:30:15Z", "updated_at": "2025-05-28T15:04:09Z", "html_url": "https://github.com/huggingface/peft/issues/2501", "labels": [], "model": "Llama", "trainer": "trl", "specialties": [ "bitsandbytes", "4bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2502, "title": "After using peft_model.merge_and_unload(), the merged model is completely different from the inference result of peft_model.", "state": "closed", "author": "weiminw", "created_at": "2025-04-17T03:36:36Z", "updated_at": "2025-05-25T15:03:46Z", "html_url": "https://github.com/huggingface/peft/issues/2502", "labels": [], "model": "unknown", "trainer": "unsloth", "specialties": [ "bitsandbytes", "xformers" ], "training_type": "inference", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "high", "training_type": "medium", "peft_method": "low", "experience_score": "medium" } }, { "number": 1831, "title": "RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu! (when checking argument for argument mat2 in method wrapper_CUDA_mm)", "state": "closed", "author": "kevalshah90", "created_at": "2024-06-05T18:53:16Z", "updated_at": "2025-05-23T00:50:05Z", "html_url": "https://github.com/huggingface/peft/issues/1831", "labels": [], "model": "Mistral", "trainer": "unknown", "specialties": [ "bitsandbytes", "4bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2445, "title": "Adding support for Conv1d", "state": "closed", "author": "EskildAndersen", "created_at": "2025-03-21T08:49:24Z", "updated_at": "2025-05-21T22:24:32Z", "html_url": "https://github.com/huggingface/peft/issues/2445", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "DoRA", "experience_score": 7, "experience_rationale": "Issue has clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 1654, "title": "Integrating ReFT", "state": "closed", "author": "raven38", "created_at": "2024-04-15T14:46:11Z", "updated_at": "2025-05-19T03:30:47Z", "html_url": "https://github.com/huggingface/peft/issues/1654", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "ReFT", "experience_score": 6, "experience_rationale": "Issue has detailed description.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2535, "title": "Error when loading multiple lora in Qwenvl-2.5-3B Model", "state": "closed", "author": "benjamintam0607", "created_at": "2025-05-08T06:39:34Z", "updated_at": "2025-05-12T01:43:25Z", "html_url": "https://github.com/huggingface/peft/issues/2535", "labels": [], "model": "Qwen", "trainer": "llamafactory", "specialties": [ "flash_attention", "4bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 453, "title": "The weights saved during training using deepspeed zero stage3 are incomplete.", "state": "closed", "author": "bestpredicts", "created_at": "2023-05-16T03:35:11Z", "updated_at": "2025-05-11T22:22:48Z", "html_url": "https://github.com/huggingface/peft/issues/453", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2441, "title": "Integarting RandLoRA a full rank PEFT algorithm", "state": "closed", "author": "PaulAlbert31", "created_at": "2025-03-19T23:28:25Z", "updated_at": "2025-05-09T13:52:08Z", "html_url": "https://github.com/huggingface/peft/issues/2441", "labels": [ "wip" ], "model": "CLIP", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Issue has detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 761, "title": "fine-tuning OpenClip with Hugingface's PEFT (such as LoRA) ", "state": "closed", "author": "KyanChen", "created_at": "2023-07-28T09:44:35Z", "updated_at": "2025-05-06T06:11:15Z", "html_url": "https://github.com/huggingface/peft/issues/761", "labels": [ "wip" ], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Issue has clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2452, "title": "Gemma with softprompt raises error", "state": "closed", "author": "krishnakanthnakkav2", "created_at": "2025-03-26T10:48:01Z", "updated_at": "2025-05-05T12:22:28Z", "html_url": "https://github.com/huggingface/peft/issues/2452", "labels": [], "model": "Gemma", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "Prompt Tuning", "experience_score": 10, "experience_rationale": "Issue shows error details provided, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2528, "title": "Finetuning of CLIP-2 with LoRA", "state": "closed", "author": "d-graz", "created_at": "2025-05-02T00:50:03Z", "updated_at": "2025-05-02T16:02:53Z", "html_url": "https://github.com/huggingface/peft/issues/2528", "labels": [], "model": "BLIP", "trainer": "unknown", "specialties": [ "bitsandbytes", "4bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "high", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2426, "title": "PEFT LoRA forwarding latency mismatch", "state": "closed", "author": "leexcape", "created_at": "2025-03-12T18:42:55Z", "updated_at": "2025-04-30T15:04:27Z", "html_url": "https://github.com/huggingface/peft/issues/2426", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2508, "title": "How to save the custom module into adapter_model.safetensrs when integrating new peft method", "state": "closed", "author": "AaronZLT", "created_at": "2025-04-22T15:46:39Z", "updated_at": "2025-04-30T11:01:58Z", "html_url": "https://github.com/huggingface/peft/issues/2508", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 7, "experience_rationale": "Issue has clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2520, "title": "Loading LoRA adapters with PeftModel.from_pretrained leads to inconsistent weights across GPUs", "state": "closed", "author": "ahmedZguir", "created_at": "2025-04-28T11:16:39Z", "updated_at": "2025-04-30T08:26:56Z", "html_url": "https://github.com/huggingface/peft/issues/2520", "labels": [], "model": "Llama", "trainer": "unknown", "specialties": [], "training_type": "inference", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "medium", "peft_method": "low", "experience_score": "medium" } }, { "number": 2505, "title": "QLoRA + FSDP bugs", "state": "closed", "author": "komaksym", "created_at": "2025-04-22T00:53:55Z", "updated_at": "2025-04-23T23:32:59Z", "html_url": "https://github.com/huggingface/peft/issues/2505", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [ "DeepSpeed", "FSDP", "bitsandbytes", "4bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2425, "title": "[Feature Request] LoRA-Null: Low-Rank Adaptation via Null Space for Large Language Models", "state": "closed", "author": "sailfish009", "created_at": "2025-03-12T10:37:46Z", "updated_at": "2025-04-20T15:03:33Z", "html_url": "https://github.com/huggingface/peft/issues/2425", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2500, "title": "Apply Zero-3 and LoRA appears empty lora weight [0]", "state": "closed", "author": "jiangxinke", "created_at": "2025-04-16T10:50:07Z", "updated_at": "2025-04-19T09:37:31Z", "html_url": "https://github.com/huggingface/peft/issues/2500", "labels": [], "model": "Qwen", "trainer": "unknown", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "high", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2438, "title": "Loading adapters in training mode produces incorrect results and randomly switches back to base model", "state": "closed", "author": "sei-cabidi", "created_at": "2025-03-19T14:37:41Z", "updated_at": "2025-04-18T15:07:39Z", "html_url": "https://github.com/huggingface/peft/issues/2438", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2494, "title": "modules_to_save not working in add_adapter() with PEFT + Diffusers", "state": "closed", "author": "zzhang2816", "created_at": "2025-04-12T03:07:15Z", "updated_at": "2025-04-17T11:51:51Z", "html_url": "https://github.com/huggingface/peft/issues/2494", "labels": [], "model": "CogVideo", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2477, "title": "Bug with 'PromptEncoderConfig' in Newer Versions of PEFT", "state": "closed", "author": "nikiduki", "created_at": "2025-04-02T13:52:18Z", "updated_at": "2025-04-17T10:25:07Z", "html_url": "https://github.com/huggingface/peft/issues/2477", "labels": [ "bug" ], "model": "BERT", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 532, "title": "RuntimeError: The size of tensor a (2048) must match the size of tensor b (2250) at non-singleton dimension 3", "state": "closed", "author": "ban1989ban", "created_at": "2023-06-01T17:10:56Z", "updated_at": "2025-04-15T01:37:23Z", "html_url": "https://github.com/huggingface/peft/issues/532", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [ "bitsandbytes", "8bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2413, "title": "`LoraConfig` multiple properties should be unified", "state": "closed", "author": "Qubitium", "created_at": "2025-03-07T04:14:24Z", "updated_at": "2025-04-14T15:04:14Z", "html_url": "https://github.com/huggingface/peft/issues/2413", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "vLLM" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2495, "title": "peft", "state": "closed", "author": "flyingdevifish", "created_at": "2025-04-12T13:01:37Z", "updated_at": "2025-04-14T12:30:43Z", "html_url": "https://github.com/huggingface/peft/issues/2495", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "Issue shows error details provided, detailed description, system info provided.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2410, "title": "running forward loop using get_peft_model disables requires_grad on output", "state": "closed", "author": "Hamidreza3252", "created_at": "2025-03-06T05:12:42Z", "updated_at": "2025-04-13T15:03:40Z", "html_url": "https://github.com/huggingface/peft/issues/2410", "labels": [], "model": "Qwen", "trainer": "peft_direct", "specialties": [ "8bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Issue has detailed description, clear title.", "confidence": { "model": "medium", "trainer_framework": "medium", "specialties": "high", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 1583, "title": "cannot import name 'prepare_model_for_int8_training' from 'peft'", "state": "closed", "author": "Eyict", "created_at": "2024-03-24T07:44:42Z", "updated_at": "2025-04-12T12:59:42Z", "html_url": "https://github.com/huggingface/peft/issues/1583", "labels": [], "model": "Llama", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 7, "experience_rationale": "Issue has clear title.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2485, "title": "X-LoRA keyerror if expert project not exists", "state": "closed", "author": "yanyongyu", "created_at": "2025-04-09T12:46:26Z", "updated_at": "2025-04-11T09:00:36Z", "html_url": "https://github.com/huggingface/peft/issues/2485", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "XLora", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium" } }, { "number": 2405, "title": "SafetensorError when Merging LoRA Weights", "state": "closed", "author": "mianwu01", "created_at": "2025-03-03T05:22:05Z", "updated_at": "2025-04-10T15:04:22Z", "html_url": "https://github.com/huggingface/peft/issues/2405", "labels": [], "model": "Llama", "trainer": "DPOTrainer", "specialties": [ "bitsandbytes" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2486, "title": "LoRA Bias Merge Not Done Correctly", "state": "closed", "author": "petercall", "created_at": "2025-04-09T18:44:34Z", "updated_at": "2025-04-10T14:19:24Z", "html_url": "https://github.com/huggingface/peft/issues/2486", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows detailed description, clear title, system info provided.", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2469, "title": "[Feature Request] Add LoRA-FA to PEFT", "state": "closed", "author": "AaronZLT", "created_at": "2025-03-29T13:23:33Z", "updated_at": "2025-04-10T12:29:40Z", "html_url": "https://github.com/huggingface/peft/issues/2469", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "Issue has detailed description, clear title.", "confidence": { "model": "low", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 2388, "title": "ValueError: Target module Qwen2_5_VisionTransformerPretrainedModel is not supported.", "state": "closed", "author": "samuellimabraz", "created_at": "2025-02-19T15:09:17Z", "updated_at": "2025-04-09T16:23:53Z", "html_url": "https://github.com/huggingface/peft/issues/2388", "labels": [], "model": "Qwen", "trainer": "transformers.Trainer", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "high", "specialties": "low", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2318, "title": "Issue merging a Lora model to a SANA transformer", "state": "closed", "author": "frutiemax92", "created_at": "2025-01-10T01:24:35Z", "updated_at": "2025-04-09T15:05:41Z", "html_url": "https://github.com/huggingface/peft/issues/2318", "labels": [], "model": "SANA", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, detailed description, clear title.", "confidence": { "model": "high", "trainer_framework": "medium", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium" } }, { "number": 137, "title": "RuntimeError: element 0 of tensors does not require grad and does not have a grad_fn", "state": "closed", "author": "PetroMaslov", "created_at": "2023-02-26T05:35:13Z", "updated_at": "2025-04-09T09:29:53Z", "html_url": "https://github.com/huggingface/peft/issues/137", "labels": [ "solved" ], "model": "T5", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "Issue shows includes code snippet, error details provided, detailed description.", "confidence": { "model": "medium", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "medium", "experience_score": "medium" } }, { "number": 2381, "title": "Bug when deleting adapters of a model with modules_to_save", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-02-17T11:22:34Z", "updated_at": "2025-04-08T14:46:23Z", "html_url": "https://github.com/huggingface/peft/issues/2381", "labels": [ "bug", "wip" ], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2394, "title": "TP + DP training error", "state": "closed", "author": "iMountTai", "created_at": "2025-02-24T08:30:53Z", "updated_at": "2025-04-04T15:04:04Z", "html_url": "https://github.com/huggingface/peft/issues/2394", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "TensorParallel" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2470, "title": "Potential bug when training LoRA with trainable tokens", "state": "closed", "author": "bil-y", "created_at": "2025-03-31T13:07:19Z", "updated_at": "2025-04-01T10:03:59Z", "html_url": "https://github.com/huggingface/peft/issues/2470", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2292, "title": "Cannot import name 'EncoderDecoderCache' from 'transformers'", "state": "closed", "author": "Huang-jia-xuan", "created_at": "2024-12-21T09:00:04Z", "updated_at": "2025-03-31T06:50:20Z", "html_url": "https://github.com/huggingface/peft/issues/2292", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2450, "title": "modules_to_save resulting in empty tensor with deepspeed zero3 LoRA training", "state": "closed", "author": "agokrani", "created_at": "2025-03-25T15:03:45Z", "updated_at": "2025-03-27T13:08:11Z", "html_url": "https://github.com/huggingface/peft/issues/2450", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1486, "title": "Save label information for seq/token classification", "state": "closed", "author": "nbroad1881", "created_at": "2024-02-19T17:12:54Z", "updated_at": "2025-03-26T15:00:35Z", "html_url": "https://github.com/huggingface/peft/issues/1486", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1967, "title": "MobileViT does not work with Inference with different LoRA adapters in the same batch", "state": "closed", "author": "saeid93", "created_at": "2024-07-29T12:53:39Z", "updated_at": "2025-03-25T15:04:29Z", "html_url": "https://github.com/huggingface/peft/issues/1967", "labels": [], "model": "MobileViT", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes environment details", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2377, "title": "Contributing new model merging method to PEFT", "state": "closed", "author": "yuang-lee", "created_at": "2025-02-14T12:17:46Z", "updated_at": "2025-03-24T15:04:11Z", "html_url": "https://github.com/huggingface/peft/issues/2377", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 7, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2407, "title": "RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cuda:3! (when checking argument for argument mat2 in method wrapper_CUDA_mm)", "state": "closed", "author": "maxliang114514", "created_at": "2025-03-04T18:09:43Z", "updated_at": "2025-03-21T16:38:44Z", "html_url": "https://github.com/huggingface/peft/issues/2407", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "VeRA", "experience_score": 10, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2281, "title": "Incompatibility of X-LoRA and MistralForSequenceClassification", "state": "closed", "author": "cyx96", "created_at": "2024-12-13T08:42:00Z", "updated_at": "2025-03-21T15:04:19Z", "html_url": "https://github.com/huggingface/peft/issues/2281", "labels": [], "model": "Mistral", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "XLora", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1846, "title": "`Conv1D` not supported using AdaLora", "state": "closed", "author": "bukun46", "created_at": "2024-06-11T16:00:32Z", "updated_at": "2025-03-21T08:50:48Z", "html_url": "https://github.com/huggingface/peft/issues/1846", "labels": [], "model": "Whisper", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "AdaLoRA", "experience_score": 10, "experience_rationale": "includes error traceback, clear title with detailed description", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2329, "title": "Request to intergrate Structure Sparsity-based PEFT (S2FT)", "state": "closed", "author": "Hanyuezhuohua", "created_at": "2025-01-14T22:18:53Z", "updated_at": "2025-03-20T15:04:19Z", "html_url": "https://github.com/huggingface/peft/issues/2329", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 9, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2421, "title": "Adapters saved in float16 are loaded in float32", "state": "closed", "author": "BenjVig", "created_at": "2025-03-11T17:39:43Z", "updated_at": "2025-03-20T10:42:06Z", "html_url": "https://github.com/huggingface/peft/issues/2421", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [ "quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2437, "title": "LoRA merge and unmerge adapters feature", "state": "closed", "author": "DavidePaglieri", "created_at": "2025-03-19T14:16:23Z", "updated_at": "2025-03-19T16:11:40Z", "html_url": "https://github.com/huggingface/peft/issues/2437", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "vLLM" ], "training_type": "RL", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2422, "title": "load_adapter Fails When modules_to_save Are Different for Each Adapter", "state": "closed", "author": "saeid93", "created_at": "2025-03-11T19:53:40Z", "updated_at": "2025-03-19T09:58:00Z", "html_url": "https://github.com/huggingface/peft/issues/2422", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1381, "title": "wrong model loading after merge_and_unload + save_pretrained", "state": "closed", "author": "sd3ntato", "created_at": "2024-01-22T13:23:21Z", "updated_at": "2025-03-19T09:06:53Z", "html_url": "https://github.com/huggingface/peft/issues/1381", "labels": [], "model": "unknown", "trainer": "trl", "specialties": [ "bitsandbytes", "4bit_quantization", "AWQ" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 692, "title": "Calling `merge_and_unload` then `save_pretrained` uploads weights twice", "state": "closed", "author": "fozziethebeat", "created_at": "2023-07-13T09:22:30Z", "updated_at": "2025-03-19T09:06:40Z", "html_url": "https://github.com/huggingface/peft/issues/692", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2429, "title": "Adapter_config is saved incorrectly during IA3 training", "state": "closed", "author": "KQDtianxiaK", "created_at": "2025-03-14T05:57:09Z", "updated_at": "2025-03-17T15:31:10Z", "html_url": "https://github.com/huggingface/peft/issues/2429", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "IA3", "experience_score": 10, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1960, "title": "Inference with different LoRA adapters in the same batch does not use the correct module_to_save classifier", "state": "closed", "author": "saeid93", "created_at": "2024-07-26T09:58:16Z", "updated_at": "2025-03-17T10:11:54Z", "html_url": "https://github.com/huggingface/peft/issues/1960", "labels": [ "wip", "contributions-welcome" ], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "inference", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes environment details, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2424, "title": "UserWarning: MatMul8bitLt: inputs will be cast from torch.float32 to float16 during quantization", "state": "closed", "author": "suhyun01150", "created_at": "2025-03-12T05:54:01Z", "updated_at": "2025-03-17T03:22:27Z", "html_url": "https://github.com/huggingface/peft/issues/2424", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "bitsandbytes", "4bit_quantization", "8bit_quantization" ], "training_type": "unsure", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2379, "title": "prompt_tuning_peft tutorial raises cache layer error", "state": "closed", "author": "jakerobers", "created_at": "2025-02-15T00:10:11Z", "updated_at": "2025-03-16T02:35:40Z", "html_url": "https://github.com/huggingface/peft/issues/2379", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [], "training_type": "supervised", "peft_method": "Prompt Tuning", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2359, "title": "Inconsistent documentation", "state": "closed", "author": "makelinux", "created_at": "2025-02-04T07:25:29Z", "updated_at": "2025-03-14T15:04:07Z", "html_url": "https://github.com/huggingface/peft/issues/2359", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 7, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2415, "title": "size mismatch for lm_head when fintune QWEN2.5", "state": "closed", "author": "minmie", "created_at": "2025-03-10T02:45:29Z", "updated_at": "2025-03-14T13:17:33Z", "html_url": "https://github.com/huggingface/peft/issues/2415", "labels": [], "model": "Qwen", "trainer": "SFTTrainer", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 636, "title": "How to save full model weights and not just the adapters ?", "state": "closed", "author": "azayz", "created_at": "2023-06-26T15:30:48Z", "updated_at": "2025-03-13T11:52:23Z", "html_url": "https://github.com/huggingface/peft/issues/636", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "inference", "peft_method": "unknown", "experience_score": 6, "experience_rationale": "includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2069, "title": "Unaligned blit request with RoBERTa", "state": "closed", "author": "vrmer", "created_at": "2024-09-14T18:10:13Z", "updated_at": "2025-03-13T01:20:42Z", "html_url": "https://github.com/huggingface/peft/issues/2069", "labels": [], "model": "RoBERTa", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2336, "title": "After using peft, the performance indicators decreased.", "state": "closed", "author": "KQDtianxiaK", "created_at": "2025-01-20T17:04:33Z", "updated_at": "2025-03-12T14:49:12Z", "html_url": "https://github.com/huggingface/peft/issues/2336", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [], "training_type": "supervised", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2412, "title": "Lora_B weight becomes 0 when using AuotModel", "state": "closed", "author": "makcedward", "created_at": "2025-03-06T19:45:29Z", "updated_at": "2025-03-11T23:53:15Z", "html_url": "https://github.com/huggingface/peft/issues/2412", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2111, "title": "could not finetune gemma 2 9b with lora and fsdp ", "state": "closed", "author": "imadoualid", "created_at": "2024-09-29T13:29:56Z", "updated_at": "2025-03-11T02:55:06Z", "html_url": "https://github.com/huggingface/peft/issues/2111", "labels": [], "model": "Gemma", "trainer": "SFTTrainer", "specialties": [ "FSDP", "bitsandbytes", "4bit_quantization", "flash_attention" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2354, "title": "Commented PeftConfig", "state": "closed", "author": "moghadas76", "created_at": "2025-01-31T11:33:50Z", "updated_at": "2025-03-10T15:04:20Z", "html_url": "https://github.com/huggingface/peft/issues/2354", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2355, "title": "dataclass config handling", "state": "closed", "author": "moghadas76", "created_at": "2025-01-31T14:48:29Z", "updated_at": "2025-03-10T15:04:18Z", "html_url": "https://github.com/huggingface/peft/issues/2355", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "4bit_quantization" ], "training_type": "unsure", "peft_method": "unknown", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1972, "title": "PEFT QLoRA with DeepSpeed ZeRO3 on multiple GPUs not working as in documentation ", "state": "closed", "author": "echo-yi", "created_at": "2024-07-30T01:33:55Z", "updated_at": "2025-03-07T01:23:31Z", "html_url": "https://github.com/huggingface/peft/issues/1972", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [ "DeepSpeed", "bitsandbytes", "4bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2390, "title": "Bug: Using 2 LoRA configs with `target_modules='all-linear'` leads to nested LoRA layers", "state": "closed", "author": "BenjaminBossan", "created_at": "2025-02-20T12:34:35Z", "updated_at": "2025-03-04T16:16:16Z", "html_url": "https://github.com/huggingface/peft/issues/2390", "labels": [ "bug", "good first issue" ], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2339, "title": "Peft version upgrade from 0.4.0 to 0.14.0 results in \"No module named \\u0027peft.utils.config\\u0027\" error", "state": "closed", "author": "incchar", "created_at": "2025-01-21T20:00:07Z", "updated_at": "2025-03-02T15:03:46Z", "html_url": "https://github.com/huggingface/peft/issues/2339", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "inference", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "includes environment details, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1402, "title": "LoRA adapters weights remain zero after training with deepspeed", "state": "closed", "author": "schwarzwalder93", "created_at": "2024-01-27T08:30:35Z", "updated_at": "2025-03-01T14:43:51Z", "html_url": "https://github.com/huggingface/peft/issues/1402", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1349, "title": "'PeftModelForCausalLM' object has no attribute 'base_model' for a deepspeed stage 3 model", "state": "closed", "author": "olegsinavski", "created_at": "2024-01-11T14:00:34Z", "updated_at": "2025-02-27T09:30:02Z", "html_url": "https://github.com/huggingface/peft/issues/1349", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [ "DeepSpeed" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2400, "title": "processing_class and tokenizer arguments on SFTTrainer()", "state": "closed", "author": "ErikKankaTrea", "created_at": "2025-02-26T12:48:33Z", "updated_at": "2025-02-27T03:39:02Z", "html_url": "https://github.com/huggingface/peft/issues/2400", "labels": [], "model": "unknown", "trainer": "SFTTrainer", "specialties": [ "DeepSpeed" ], "training_type": "supervised", "peft_method": "unknown", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2275, "title": "TypeError: LoraConfig.__init__() got an unexpected keyword argument 'eva_config'", "state": "closed", "author": "Mohankrish08", "created_at": "2024-12-11T18:45:42Z", "updated_at": "2025-02-26T10:58:54Z", "html_url": "https://github.com/huggingface/peft/issues/2275", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1035, "title": "Quality of inference severely changed after merge_and_unload", "state": "closed", "author": "blaze7451", "created_at": "2023-10-18T18:57:53Z", "updated_at": "2025-02-25T21:24:53Z", "html_url": "https://github.com/huggingface/peft/issues/1035", "labels": [], "model": "Llama", "trainer": "none", "specialties": [], "training_type": "inference", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes environment details", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2315, "title": "Prefix Tuning dimension error with Qwen2 and missing vocab_size for PaliGemma2", "state": "closed", "author": "Florian-Dreyer", "created_at": "2025-01-08T22:52:17Z", "updated_at": "2025-02-25T15:04:15Z", "html_url": "https://github.com/huggingface/peft/issues/2315", "labels": [], "model": "Qwen", "trainer": "peft_direct", "specialties": [ "DeepSpeed" ], "training_type": "unsure", "peft_method": "Prefix Tuning", "experience_score": 10, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 244, "title": "Implementation of adapter?", "state": "closed", "author": "jzhang38", "created_at": "2023-03-31T04:42:41Z", "updated_at": "2025-02-24T15:04:29Z", "html_url": "https://github.com/huggingface/peft/issues/244", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 6, "experience_rationale": "moderate detail provided", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2330, "title": "MoELorA", "state": "closed", "author": "moghadas76", "created_at": "2025-01-15T09:29:58Z", "updated_at": "2025-02-23T15:03:30Z", "html_url": "https://github.com/huggingface/peft/issues/2330", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "MoELoRA", "experience_score": 6, "experience_rationale": "moderate detail provided", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2208, "title": "TypeError: LoraConfig.__init__() got an unexpected keyword argument 'exclude_modules'", "state": "closed", "author": "Imran-tech-lab", "created_at": "2024-11-09T13:57:02Z", "updated_at": "2025-02-20T07:06:49Z", "html_url": "https://github.com/huggingface/peft/issues/2208", "labels": [], "model": "Qwen", "trainer": "none", "specialties": [ "bitsandbytes" ], "training_type": "inference", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 71, "title": "Use `peft` for RLHF", "state": "closed", "author": "younesbelkada", "created_at": "2023-02-11T08:37:44Z", "updated_at": "2025-02-19T11:55:53Z", "html_url": "https://github.com/huggingface/peft/issues/71", "labels": [ "wip" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "RL", "peft_method": "unknown", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2273, "title": "Support for Custom Adapters", "state": "closed", "author": "dgme-syz", "created_at": "2024-12-11T06:28:08Z", "updated_at": "2025-02-18T10:05:30Z", "html_url": "https://github.com/huggingface/peft/issues/2273", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2307, "title": "The provided `peft_type` 'PROMPT_TUNING' is not compatible with the `PeftMixedModel`.", "state": "closed", "author": "Radu1999", "created_at": "2025-01-07T01:46:18Z", "updated_at": "2025-02-14T15:04:03Z", "html_url": "https://github.com/huggingface/peft/issues/2307", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "Prompt Tuning", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2368, "title": "[FSDP] After training embed_tokens in modules_to_save model has hallucinations", "state": "closed", "author": "DmitryDiTy", "created_at": "2025-02-07T13:23:07Z", "updated_at": "2025-02-14T08:23:35Z", "html_url": "https://github.com/huggingface/peft/issues/2368", "labels": [], "model": "unknown", "trainer": "SFTTrainer", "specialties": [ "FSDP" ], "training_type": "supervised", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2304, "title": "a question about input_ids and attention_mask after prefix-tuning", "state": "closed", "author": "MaTengSYSU", "created_at": "2025-01-06T08:44:01Z", "updated_at": "2025-02-13T15:04:14Z", "html_url": "https://github.com/huggingface/peft/issues/2304", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [ "4bit_quantization", "8bit_quantization" ], "training_type": "unsure", "peft_method": "Prefix Tuning", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 334, "title": "Adding additional tokens to vocabulary", "state": "closed", "author": "adibMosharrof", "created_at": "2023-04-19T02:24:21Z", "updated_at": "2025-02-11T10:28:59Z", "html_url": "https://github.com/huggingface/peft/issues/334", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "inference", "peft_method": "unknown", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2367, "title": "Some weights of MistralForSequenceClassification were not initialized from the model checkpoint at mistralai/Mistral-7B-Instruct-v0.3 and are newly initialized: ['score.weight']", "state": "closed", "author": "amritansh6", "created_at": "2025-02-07T12:29:22Z", "updated_at": "2025-02-10T11:01:57Z", "html_url": "https://github.com/huggingface/peft/issues/2367", "labels": [], "model": "Mistral", "trainer": "transformers.Trainer", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2317, "title": "Issue with finetuning with Corda", "state": "closed", "author": "sirluk", "created_at": "2025-01-09T07:12:18Z", "updated_at": "2025-02-10T10:22:03Z", "html_url": "https://github.com/huggingface/peft/issues/2317", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [], "training_type": "supervised", "peft_method": "Corda", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2302, "title": "Bug in `get_peft_model_state_dict` when using vblora", "state": "closed", "author": "KaiyangLi1992", "created_at": "2024-12-31T16:52:25Z", "updated_at": "2025-02-09T15:03:35Z", "html_url": "https://github.com/huggingface/peft/issues/2302", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "unsure", "peft_method": "VBLoRA", "experience_score": 9, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2132, "title": "Key mismatch when trying to load a LORA adapter into an XLORA model", "state": "closed", "author": "p4arth", "created_at": "2024-10-05T17:03:30Z", "updated_at": "2025-02-07T15:04:04Z", "html_url": "https://github.com/huggingface/peft/issues/2132", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [], "training_type": "supervised", "peft_method": "XLora", "experience_score": 10, "experience_rationale": "includes code snippets, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2185, "title": "Xlora cannot reload model from last checkpoint by using trainer.train(resume_from_checkpoint=\"checkpp\")", "state": "closed", "author": "SongHanKen", "created_at": "2024-10-29T02:54:22Z", "updated_at": "2025-02-07T15:04:01Z", "html_url": "https://github.com/huggingface/peft/issues/2185", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [ "FSDP" ], "training_type": "supervised", "peft_method": "XLora", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2301, "title": "How to pass in an attention _ mask that is one dimension more than input _ ids", "state": "closed", "author": "Chinesehou97", "created_at": "2024-12-31T02:26:14Z", "updated_at": "2025-02-07T15:03:57Z", "html_url": "https://github.com/huggingface/peft/issues/2301", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 7, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2364, "title": "docs: broken links to boft", "state": "closed", "author": "makelinux", "created_at": "2025-02-06T14:48:16Z", "updated_at": "2025-02-07T10:14:44Z", "html_url": "https://github.com/huggingface/peft/issues/2364", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 7, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2362, "title": "Import error", "state": "closed", "author": "ikamensh", "created_at": "2025-02-05T20:19:35Z", "updated_at": "2025-02-05T20:38:50Z", "html_url": "https://github.com/huggingface/peft/issues/2362", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [ "bitsandbytes" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2296, "title": "Error of load_adapter of Target module is not supported when using Qwen2-VL", "state": "closed", "author": "hcsgxl-bot", "created_at": "2024-12-24T01:46:09Z", "updated_at": "2025-02-05T15:04:03Z", "html_url": "https://github.com/huggingface/peft/issues/2296", "labels": [], "model": "Qwen", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2348, "title": "Incorrect Magnitude Calculation for DoRA Linear Layers (Violates DoRA Paper Methodology)", "state": "closed", "author": "arcteryox", "created_at": "2025-01-26T19:43:50Z", "updated_at": "2025-01-30T18:56:52Z", "html_url": "https://github.com/huggingface/peft/issues/2348", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "DoRA", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 184, "title": "How do I continue training with PEFT?", "state": "closed", "author": "therealadityashankar", "created_at": "2023-03-17T10:12:16Z", "updated_at": "2025-01-29T21:30:38Z", "html_url": "https://github.com/huggingface/peft/issues/184", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "unknown", "experience_score": 7, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2241, "title": "Request for adding the lora implementation for Conv1d rather than transormers.utils.Conv1d", "state": "closed", "author": "HelloWorldLTY", "created_at": "2024-11-28T19:02:45Z", "updated_at": "2025-01-27T10:48:21Z", "html_url": "https://github.com/huggingface/peft/issues/2241", "labels": [ "contributions-welcome" ], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2212, "title": "Documentation for LoRAConfig.", "state": "closed", "author": "brynhayder", "created_at": "2024-11-12T16:06:40Z", "updated_at": "2025-01-27T10:08:57Z", "html_url": "https://github.com/huggingface/peft/issues/2212", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2285, "title": "TypeError: TorchaoLoraLinear.__init__() missing 1 required keyword-only argument: 'get_apply_tensor_subclass'", "state": "closed", "author": "spezialspezial", "created_at": "2024-12-17T01:32:07Z", "updated_at": "2025-01-23T07:46:46Z", "html_url": "https://github.com/huggingface/peft/issues/2285", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "torchao" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2337, "title": "AdaLora kthvalue(): selected number k out of range for dimension 0", "state": "closed", "author": "PKaralupov", "created_at": "2025-01-20T21:56:43Z", "updated_at": "2025-01-23T05:25:02Z", "html_url": "https://github.com/huggingface/peft/issues/2337", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [ "bitsandbytes", "8bit_quantization" ], "training_type": "supervised", "peft_method": "AdaLoRA", "experience_score": 10, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2283, "title": "TypeError when inference with different LoRA adapters in the same batch", "state": "closed", "author": "yuxiang-guo", "created_at": "2024-12-15T13:07:34Z", "updated_at": "2025-01-22T15:03:50Z", "html_url": "https://github.com/huggingface/peft/issues/2283", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "inference", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1962, "title": "Deprecation: Transformers will no longer support `past_key_values` to be tuples", "state": "closed", "author": "BenjaminBossan", "created_at": "2024-07-26T10:43:54Z", "updated_at": "2025-01-21T11:15:15Z", "html_url": "https://github.com/huggingface/peft/issues/1962", "labels": [ "help wanted", "wip", "contributions-welcome" ], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "Prefix Tuning", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2289, "title": "Inconsistent Parameter Mismatches After Merging PEFT and Base Models", "state": "closed", "author": "enhulu-ms", "created_at": "2024-12-19T00:54:08Z", "updated_at": "2025-01-20T17:28:17Z", "html_url": "https://github.com/huggingface/peft/issues/2289", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2326, "title": "AttributeError: ModulesToSaveWrapper has no attribute `dense`", "state": "closed", "author": "KQDtianxiaK", "created_at": "2025-01-13T16:49:37Z", "updated_at": "2025-01-20T16:29:05Z", "html_url": "https://github.com/huggingface/peft/issues/2326", "labels": [], "model": "unknown", "trainer": "none", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2277, "title": "Request to intergrate Monarch-based PEFT (MoRe)", "state": "closed", "author": "Edenzzzz", "created_at": "2024-12-13T02:03:03Z", "updated_at": "2025-01-20T15:04:03Z", "html_url": "https://github.com/huggingface/peft/issues/2277", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "unknown", "experience_score": 9, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2278, "title": " Adding Dynamic Low-Rank Adaptation (DoRA ACL2024) ", "state": "closed", "author": "dohuyduc2002", "created_at": "2024-12-13T04:23:12Z", "updated_at": "2025-01-20T15:04:01Z", "html_url": "https://github.com/huggingface/peft/issues/2278", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "DoRA", "experience_score": 9, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 114, "title": "AttributeError: 'DataParallel' object has no attribute 'config' when call get_peft_model.", "state": "closed", "author": "hjsg1010", "created_at": "2023-02-20T09:49:54Z", "updated_at": "2025-01-19T16:47:15Z", "html_url": "https://github.com/huggingface/peft/issues/114", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [ "DataParallel" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2262, "title": "Could you provide example code for AdaLoRA finetuning decoder-only model?", "state": "closed", "author": "yuang-lee", "created_at": "2024-12-05T12:03:31Z", "updated_at": "2025-01-18T15:03:29Z", "html_url": "https://github.com/huggingface/peft/issues/2262", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "supervised", "peft_method": "AdaLoRA", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2270, "title": "Different Results When Predicting with Multiple LoRA Adapters in a Loop VS. Using only One LoRA", "state": "closed", "author": "beyondguo", "created_at": "2024-12-10T06:53:14Z", "updated_at": "2025-01-18T15:03:27Z", "html_url": "https://github.com/huggingface/peft/issues/2270", "labels": [], "model": "unknown", "trainer": "transformers.Trainer", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2266, "title": "Can't PromptTuning in Multi-GPU with DeepSpeed and Qwen2.5-14B-Instruct", "state": "closed", "author": "dongshou", "created_at": "2024-12-09T11:13:06Z", "updated_at": "2025-01-17T15:03:49Z", "html_url": "https://github.com/huggingface/peft/issues/2266", "labels": [], "model": "Qwen", "trainer": "transformers.Trainer", "specialties": [ "DeepSpeed", "flash_attention" ], "training_type": "supervised", "peft_method": "Prompt Tuning", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 832, "title": "Lora caused memory leak for simple finetuning ", "state": "closed", "author": "Randolph-zeng", "created_at": "2023-08-17T08:43:08Z", "updated_at": "2025-01-17T10:03:50Z", "html_url": "https://github.com/huggingface/peft/issues/832", "labels": [], "model": "unknown", "trainer": "SFTTrainer", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2322, "title": "model merge and unload feature for AdaLora", "state": "closed", "author": "DaehanKim", "created_at": "2025-01-12T09:20:01Z", "updated_at": "2025-01-14T12:47:35Z", "html_url": "https://github.com/huggingface/peft/issues/2322", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [ "4bit_quantization", "8bit_quantization" ], "training_type": "supervised", "peft_method": "AdaLoRA", "experience_score": 9, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2319, "title": "Import error , is it a version issue?", "state": "closed", "author": "zhangyangniubi", "created_at": "2025-01-10T02:34:52Z", "updated_at": "2025-01-13T10:13:18Z", "html_url": "https://github.com/huggingface/peft/issues/2319", "labels": [], "model": "unknown", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "unknown", "experience_score": 9, "experience_rationale": "includes environment details, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "low", "specialties": "low", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 793, "title": "Lora model (after fine tuning) working exactly the same as base model ", "state": "closed", "author": "Luke-4", "created_at": "2023-08-06T18:55:52Z", "updated_at": "2025-01-13T08:43:19Z", "html_url": "https://github.com/huggingface/peft/issues/793", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 9, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2259, "title": "Deepseek_lora custom keys in input_data fails", "state": "closed", "author": "Zhaoyi-Yan", "created_at": "2024-12-05T02:20:27Z", "updated_at": "2025-01-12T15:03:31Z", "html_url": "https://github.com/huggingface/peft/issues/2259", "labels": [], "model": "DeepSeek", "trainer": "unknown", "specialties": [], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "high", "trainer_framework": "low", "specialties": "low", "training_type": "low", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2260, "title": "Is it possible to support the transformer engine when using Lora in Megatron?", "state": "closed", "author": "liulong11", "created_at": "2024-12-05T03:24:15Z", "updated_at": "2025-01-12T15:03:29Z", "html_url": "https://github.com/huggingface/peft/issues/2260", "labels": [], "model": "unknown", "trainer": "Megatron", "specialties": [ "Megatron" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 8, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1432, "title": "ValueError: Cannot flatten integer dtype tensors", "state": "closed", "author": "fancyerii", "created_at": "2024-02-04T03:26:10Z", "updated_at": "2025-01-12T12:15:57Z", "html_url": "https://github.com/huggingface/peft/issues/1432", "labels": [], "model": "unknown", "trainer": "SFTTrainer", "specialties": [ "FSDP", "bitsandbytes", "4bit_quantization" ], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, includes environment details", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 1083, "title": "Use AutoPeftModelForCausalLM to merge and save model cause crashed in the low VRAM situation", "state": "closed", "author": "ytchen175", "created_at": "2023-11-06T06:19:36Z", "updated_at": "2025-01-11T20:30:55Z", "html_url": "https://github.com/huggingface/peft/issues/1083", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [ "4bit_quantization", "8bit_quantization" ], "training_type": "unsure", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes error traceback, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 218, "title": "Pipeline 'text-generation' support when?", "state": "closed", "author": "gururise", "created_at": "2023-03-26T18:47:06Z", "updated_at": "2025-01-11T20:01:39Z", "html_url": "https://github.com/huggingface/peft/issues/218", "labels": [ "solved" ], "model": "unknown", "trainer": "Megatron", "specialties": [ "Megatron" ], "training_type": "unsure", "peft_method": "unknown", "experience_score": 7, "experience_rationale": "clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "high", "training_type": "high", "peft_method": "low", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2250, "title": "frozen modules also be lora", "state": "closed", "author": "onehaitao", "created_at": "2024-12-03T12:26:54Z", "updated_at": "2025-01-11T15:03:43Z", "html_url": "https://github.com/huggingface/peft/issues/2250", "labels": [], "model": "unknown", "trainer": "peft_direct", "specialties": [], "training_type": "supervised", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes environment details, clear title with detailed description", "confidence": { "model": "low", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } }, { "number": 2316, "title": "peft with DinoV2 and tasktype feature extraction", "state": "closed", "author": "createdaccountbecauseIwantgithubcopilot", "created_at": "2025-01-09T02:48:36Z", "updated_at": "2025-01-09T14:11:54Z", "html_url": "https://github.com/huggingface/peft/issues/2316", "labels": [], "model": "DINOv2", "trainer": "none", "specialties": [], "training_type": "inference", "peft_method": "LoRA", "experience_score": 10, "experience_rationale": "includes code snippets, includes error traceback", "confidence": { "model": "high", "trainer_framework": "high", "specialties": "low", "training_type": "high", "peft_method": "high", "experience_score": "medium", "experience_rationale": "medium" } } ]