README.md

---
license: apache-2.0
base_model: Qwen/Qwen2.5-Coder-14B
tags:
- code
- qwen2.5
- lora-merged
- fine-tuned
library_name: transformers
---

# PoPilot - Fine-tuned Qwen2.5-Coder-14B

This model is a fine-tuned version of [Qwen/Qwen2.5-Coder-14B](https://huggingface.co/Qwen/Qwen2.5-Coder-14B) with LoRA adapters merged.

## Model Details

- **Base Model**: Qwen/Qwen2.5-Coder-14B
- **Fine-tuning Method**: LoRA (Low-Rank Adaptation)
- **Training**: Supervised Fine-Tuning (SFT)
- **Merged**: Full model weights (LoRA merged with base)

## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained(
    "Justin6657/PoPilot",
    torch_dtype="auto",
    device_map="auto",
    trust_remote_code=True
)

tokenizer = AutoTokenizer.from_pretrained(
    "Justin6657/PoPilot",
    trust_remote_code=True
)

# Example usage
prompt = "Write a Python function to calculate fibonacci numbers:"
inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_length=200, temperature=0.7)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)
```

## Training Details

This model was fine-tuned using LoRA adapters and then merged back into the full model weights.
Original LoRA checkpoint path: `/net/projects/CLS/DSI_clinic/justin/checkpoint/augmented_train_Qwen2.5-Coder-14B_full-model_repair-synth_repair-simple-phase4`