https://github.com/ise-uiuc/magicoder

[ICML'24] Magicoder: Empowering Code Generation with OSS-Instruct
https://github.com/ise-uiuc/magicoder

ai4code large-language-models llm llm4code

Last synced: about 15 hours ago
JSON representation

[ICML'24] Magicoder: Empowering Code Generation with OSS-Instruct

• Host: GitHub
• URL: https://github.com/ise-uiuc/magicoder
• Owner: ise-uiuc
• License: mit
• Created: 2023-11-10T07:35:29.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2024-11-01T17:28:52.000Z (about 2 months ago)
• Last Synced: 2024-12-14T01:01:06.162Z (8 days ago)
• Topics: ai4code, large-language-models, llm, llm4code
• Language: Python
• Homepage: https://proceedings.mlr.press/v235/wei24h.html
• Size: 2.4 MB
• Stars: 1,983
• Watchers: 25
• Forks: 166
• Open Issues: 5
• Metadata Files:
  • Readme: README-DEV.md
  • License: LICENSE
  • Citation: CITATION.cff

Awesome Lists containing this project

• StarryDivineSky - ise-uiuc/magicoder - Intit 提供支持的模型系列，这是一种新颖的方法LLMs，通过开源代码片段为代码生成低偏差和高质量的指令数据。OSS-Instruct 通过赋予LLM它们丰富的开源引用来产生更多样化、更真实和可控的数据，从而减轻了合成指令数据的固有偏见。 (A01_文本生成_文本对话 / 大语言对话模型及数据)

README

        
# Implementation Details of 🎩Magicoder

> [!WARNING]

> This documentation is still WIP. Raise an [issue](https://github.com/ise-uiuc/magicoder/issues) in case you found any errors.

## Data collection and generation

Make sure you have set up your `OPENAI_API_KEY` and optionally `OPENAI_BASE_URL`. Then run with

```bash

python src/magicoder/generate_data.py \

  --seed_code_start_index ${START_INDEX_OF_RAW_DATA} \

  --max_new_data ${MAX_DATA_TO_GENERATE} \

  --data_dir python \

  --tag python

```

To continue an interrupted run, use `--continue_from` flag:

```bash

python src/magicoder/generate_data.py \

  --seed_code_start_index ${START_INDEX_OF_RAW_DATA} \

  --max_new_data ${MAX_DATA_TO_GENERATE} \

  --data_dir python \

  --continue_from ${PATH_TO_DATA_FILE}

```

## Data cleaning and decontamination

After the data collection, clean and decontaminate the data with the following command:

```bash

python src/magicoder/clean_data.py --data_files {PATH_TO_DATA_FILE} --output_file {CLEANING_OUTPUT_PATH}

python -m magicoder.decontamination.find_substrings \

  --dataset_name "json" \

  --output_file ${DECONTAM_OUTPUT_PATH} \

  --output_dir ${OUTPUT_DIR} \

  --columns problem solution \

  --data_files ${PATH_TO_DATA_FILE}

```

You probably need to run this multiple times with different data files.

## Data preprocessing

Before instruction tuning, let's reformat the data into instruction-response pairs:

  

```bash

python src/magicoder/preprocess_data.py \

  --dataset_path json \

  --data_files ${DECONTAM_OUTPUT_PATH} \

  --output_file ${PREPROCESS_OUTPUT_PATH} \

  --key src-instruct

```

After that, you can combine all the `jsonl` files into one.

## Instruction tuning

Pointing the environment variable `CUDA_VISIBLE_DEVICES` to the GPUs you want to use, train the model with the following command to obtain Magicoder:

```bash

accelerate launch -m magicoder.train \

  --model_key $MODEL_KEY \

  --use_flash_attention True \

  --max_training_seq_length 1216 \

  --datafile_paths \

    ${PATH_TO_OSS_INSTRUCT} \

  --output_dir $MAGICODER_OUTPUT_DIR \

  --bf16 True \

  --num_train_epochs 2 \

  --per_device_train_batch_size 2 \

  --gradient_accumulation_steps 128 \

  --group_by_length False \

  --ddp_find_unused_parameters False \

  --logging_steps 1 \

  --log_level info \

  --optim adafactor \

  --max_grad_norm -1 \

  --warmup_steps 15 \

  --learning_rate 5e-5 \

  --lr_scheduler_type linear

```

To get Magicoder-S, continue the training with the following command:

```bash

accelerate launch -m magicoder.train \

  --model_key $MODEL_KEY \

  --model_name_or_path $MAGICODER_OUTPUT_DIR \

  --use_flash_attention True \

  --max_training_seq_length 1024 \

  --datafile_paths \

    ${PATH_TO_EVOL_INSTRUCT} \

  --output_dir $MAGICODER_S_OUTPUT_DIR \

  --bf16 True \

  --num_train_epochs 2 \

  --per_device_train_batch_size 2 \

  --gradient_accumulation_steps 128 \

  --group_by_length False \

  --ddp_find_unused_parameters False \

  --logging_steps 1 \

  --log_level info \

  --optim adafactor \

  --max_grad_norm -1 \

  --warmup_steps 15 \

  --learning_rate 5e-5 \

  --lr_scheduler_type linear

```