Instructions to use YanZhanPKU/dLLM-PRM-Gap-c3a-causal-lasttoken-dream7b-s43 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use YanZhanPKU/dLLM-PRM-Gap-c3a-causal-lasttoken-dream7b-s43 with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
YanZhanPKU commited on
Commit ·
0bd585d
0
Parent(s):
Publish dLLM PRM Gap release
Browse files- .gitattributes +35 -0
- README.md +55 -0
- adapter.safetensors +3 -0
- config.json +38 -0
.gitattributes
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 9 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 10 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 11 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 12 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
| 13 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 14 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 15 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 16 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 17 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 18 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 19 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 20 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 21 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 22 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 23 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 24 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 25 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 26 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 27 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 28 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
| 29 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 30 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 31 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 32 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 33 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
base_model: Dream-org/Dream-v0-Instruct-7B
|
| 4 |
+
tags:
|
| 5 |
+
- dllm-prm-gap
|
| 6 |
+
- discrete-diffusion
|
| 7 |
+
- gsm8k
|
| 8 |
+
- process-reward-model
|
| 9 |
+
- dream-7b
|
| 10 |
+
pipeline_tag: text-classification
|
| 11 |
+
library_name: peft
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
<h1 style="text-align: left; font-size: 1.6em; margin-bottom: 0.75em;">
|
| 15 |
+
<span style="color:#E69F00; font-weight:bold;">dLLM</span><span style="color:#0072B2; font-weight:bold;"> PRM</span><span style="color:#009E73; font-weight:bold;"> Gap</span> · C3A · causal last-token · seed 43
|
| 16 |
+
</h1>
|
| 17 |
+
|
| 18 |
+
<div style="text-align: left; margin-bottom: 18px;">
|
| 19 |
+
<a href="https://github.com/dLLM-PRM-Gap/dLLM-PRM-Gap" target="_blank">💻 Code</a>
|
| 20 |
+
•
|
| 21 |
+
<a href="https://huggingface.co/collections/YanZhanPKU/dllm-prm-gap" target="_blank">🤗 Collection</a>
|
| 22 |
+
•
|
| 23 |
+
<a href="https://arxiv.org/abs/2609.35472" target="_blank">📄 Paper</a>
|
| 24 |
+
</div>
|
| 25 |
+
|
| 26 |
+
## News
|
| 27 |
+
|
| 28 |
+
- **2026-09**: Accepted at **NeurIPS 2026**.
|
| 29 |
+
|
| 30 |
+
Public adapter release for **dLLM PRM Gap**. a C3 control checkpoint with the stated attention/readout protocol and seed.
|
| 31 |
+
|
| 32 |
+
> This repository contains only trainable adapter parameters and the reward head. It does not include the base model. The release configuration is recorded in `config.json`; the arXiv paper defines the release scope and citation.
|
| 33 |
+
|
| 34 |
+
## Files
|
| 35 |
+
|
| 36 |
+
- `adapter.safetensors` — compact adapter weights
|
| 37 |
+
- `config.json` — public base-model id, architecture, and provenance
|
| 38 |
+
|
| 39 |
+
## Load
|
| 40 |
+
|
| 41 |
+
```python
|
| 42 |
+
from huggingface_hub import snapshot_download
|
| 43 |
+
from prm.checkpointing import load_diffusion_prm
|
| 44 |
+
|
| 45 |
+
path = snapshot_download("YanZhanPKU/dLLM-PRM-Gap-c3a-causal-lasttoken-dream7b-s43")
|
| 46 |
+
model, report = load_diffusion_prm(
|
| 47 |
+
checkpoint=path,
|
| 48 |
+
model_path="Dream-org/Dream-v0-Instruct-7B",
|
| 49 |
+
local_files_only=False,
|
| 50 |
+
)
|
| 51 |
+
```
|
| 52 |
+
|
| 53 |
+
**Role:** C3A causal last-token ORM-protocol control.
|
| 54 |
+
|
| 55 |
+
Paper: [https://arxiv.org/abs/2609.35472](https://arxiv.org/abs/2609.35472).
|
adapter.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f6b5d5a337e1aecdcbc52cabaa97f9ffe9147f3be860b3ab48b492c9cf1e8fc
|
| 3 |
+
size 34890548
|
config.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_model": "Dream-org/Dream-v0-Instruct-7B",
|
| 3 |
+
"adapter_type": "LoRA plus reward head",
|
| 4 |
+
"attention": "causal",
|
| 5 |
+
"pool_strategy": "last_token",
|
| 6 |
+
"mask_aware_pooling": true,
|
| 7 |
+
"step_embedding": false,
|
| 8 |
+
"lora": {
|
| 9 |
+
"r": 16,
|
| 10 |
+
"alpha": 32,
|
| 11 |
+
"dropout": 0.05,
|
| 12 |
+
"target_modules": [
|
| 13 |
+
"q_proj",
|
| 14 |
+
"v_proj"
|
| 15 |
+
]
|
| 16 |
+
},
|
| 17 |
+
"release_status": "public release",
|
| 18 |
+
"checkpoint_identity_status": "exact C3 ORM-protocol control",
|
| 19 |
+
"paper_role": "C3A causal last-token ORM-protocol control",
|
| 20 |
+
"adapter_bytes": 34890548,
|
| 21 |
+
"parameter_prefixes_kept": [
|
| 22 |
+
"lora_A",
|
| 23 |
+
"lora_B",
|
| 24 |
+
"reward_head"
|
| 25 |
+
],
|
| 26 |
+
"causal": true,
|
| 27 |
+
"no_step_embed": true,
|
| 28 |
+
"no_mask_aware": false,
|
| 29 |
+
"lora_r": 16,
|
| 30 |
+
"lora_alpha": 32,
|
| 31 |
+
"lora_dropout": 0.05,
|
| 32 |
+
"step_embed_dim": 256,
|
| 33 |
+
"reward_hidden": 1024,
|
| 34 |
+
"seed": 43,
|
| 35 |
+
"max_steps": 15000,
|
| 36 |
+
"best_step": "checkpoint-selected",
|
| 37 |
+
"orm_protocol": true
|
| 38 |
+
}
|