# EditCoder

The EditCoder models are the fine-tuned models described in the following paper:

```
@inproceedings{cassano2023edit,
      title={{Can It Edit? Evaluating the Ability of Large Language Models to Follow Code Editing Instructions}}, 
      author={Federico Cassano and Luisa Li and Akul Sethi and Noah Shinn and Abby Brennan-Jones and Anton Lozhkov and Carolyn Jane Anderson and Arjun Guha},
      booktitle={The First International Workshop on Large Language Model for Code},
      year={2024},
      url={https://arxiv.org/abs/2312.12450}
}
```

This repository has several models. The root is the fine-tune of DeepSeek Coder 33B on the EditPackFT dataset. The other models are
in subdirectories. You can do this:

```bash
AutoModelForCausalLM.from_pretrained("nuprl/EditCoder", subfolder=DIR_NAME)
`````

## Prompt
The model has been trained on the following prompt format:
```
## Code Before:
{before}
## Instruction:
{instruction}
## Code After:
{after}
```

Here is a python function that can be used for formatting the prompt correctly:
```py
def edit_prompt(old, instr):
    before = f"""## Code Before:\n{old}\n"""
    instr = f"""## Instruction:\n{instr}\n"""
    after = f"""## Code After:\n"""
    return before + instr + after
```

## Training Code

We provide the full pipeline that was used for training our own edit-coder model.
The pipeline and instructions can be found on our [GitHub repository](https://github.com/nuprl/CanItEdit/tree/main/editcoder).