---
library_name: peft
license: other
base_model: Qwen/Qwen2.5-coder-3B
tags:
- generated_from_trainer
datasets: []
model-index:
- name: outputs/qwen2.5-coder-3b-lora
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
<details><summary>See axolotl config</summary>

axolotl version: `0.10.0.dev0`
```yaml
base_model: Qwen/Qwen2.5-coder-3B
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer
trust_remote_code: true
chat_template: qwen_25

adapter: qlora
lora_r: 8
lora_alpha: 32
lora_dropout: 0.05
lora_target_modules:
  - c_attn
  - c_proj
  - w1
  - w2
  - q_proj
  - v_proj
  - k_proj
  - o_proj

load_in_4bit: true
bnb_4bit_compute_dtype: float16
bnb_4bit_use_double_quant: true
bnb_4bit_quant_type: nf4

datasets:
  - path: ./datasets/generic_formatted_data.jsonl
    type: alpaca

val_set_size: 0.01
dataset_prepared_path:

sequence_len: 2048
pad_to_sequence_len: true

output_dir: ./outputs/qwen2.5-coder-3b-lora
num_epochs: 3
micro_batch_size: 2
gradient_accumulation_steps: 8
evals_per_epoch: 1
saves_per_epoch: 1
optimizer: adamw_bnb_8bit
learning_rate: 2e-5
lr_scheduler: cosine
warmup_steps: 50

gradient_checkpointing: true
fp16: true
bf16: false
tf32: true
flash_attention: true
eager_attention: false

logging_steps: 1
debug: true
wandb_project: qwen-coder
wandb_name: qwen2.5-coder-3b-lora
wandb_log_model: "false"
wandb_mode: disabled
```

</details><br>

# outputs/qwen2.5-coder-3b-lora

This model is a fine-tuned version of [Qwen/Qwen2.5-coder-3B](https://huggingface.co/Qwen/Qwen2.5-coder-3B) on the ./datasets/generic_formatted_data.jsonl dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0817

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 2
- eval_batch_size: 2
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- gradient_accumulation_steps: 8
- total_train_batch_size: 32
- total_eval_batch_size: 4
- optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 50
- training_steps: 1375
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch  | Step | Validation Loss |
|:-------------:|:------:|:----:|:---------------:|
| 1.0456        | 0.0022 | 1    | 0.9417          |
| 0.3029        | 1.0    | 459  | 0.1403          |
| 0.044         | 2.0    | 918  | 0.0817          |


### Framework versions

- PEFT 0.15.2
- Transformers 4.51.3
- Pytorch 2.6.0+cu124
- Datasets 3.5.1
- Tokenizers 0.21.1