Data

Tools

Indexes

Applications

Experiments

Open Source Science

https://github.com/albertzhaoca/e1

https://github.com/albertzhaoca/e1

Science Score: 26.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (5.7%) to scientific vocabulary
Last synced: 10 months ago · JSON representation

Repository

Basic Info
  • Host: GitHub
  • Owner: AlbertZhaoCA
  • License: apache-2.0
  • Language: Python
  • Default Branch: main
  • Size: 5.22 MB
Statistics
  • Stars: 0
  • Watchers: 0
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created about 1 year ago · Last pushed 12 months ago
Metadata Files
Readme Contributing License Code of conduct

README.md

AlgoForge: Specializing Code Generation Agents through Collaborative Reinforcement Learning

Requirements

Software Requirements

  • Python 3.9+
  • transformers>=4.51.0
  • flash-attn>=2.4.3
  • vllm>=0.8.3

1. Train the Planner Agent

  1. Start the vLLM engine Update the model name, URL and port in examples/reward_function/higher-order.py as needed, and verify the settings in examples/higher-order.yaml.

bash python -m vllm.entrypoints.openai.api_server \ --model path-to-your-model \ --host 0.0.0.0 \ --port 8000 \ --tensor-parallel-size 4 \ --trust-remote-code

  1. Launch the training script

bash bash examples/higher-order.sh

2. Train the Coder Agent

  1. Generate the dataset Run agent2_dataset.py, then update the dataset path in either:
  • examples/codegenwc.yaml
  • examples/codegenwoc.yaml
  1. Launch the training script

bash bash examples/codewoc.sh

3. Merge Checkpoint to Hugging Face Format

Once training is complete, convert your checkpoint:

bash python3 scripts/model_merger.py \ --local_dir to-the-actor-path-of-your-model-chekcpoint

Now you’re all set to run both agents end‑to‑end!

Owner

  • Name: Albert Zhao
  • Login: AlbertZhaoCA
  • Kind: user

A CS sophomore at Kean University Interest in HCI, NLP, and software architecture

GitHub Events

Total
  • Push event: 6
  • Create event: 2
Last Year
  • Push event: 6
  • Create event: 2

Dependencies

.github/workflows/tests.yml actions
  • actions/checkout v4 composite
  • actions/setup-python v5 composite
Dockerfile docker
  • nvcr.io/nvidia/pytorch 24.08-py3 build
.github/requirements-test.txt pypi
  • codetiming * test
  • datasets * test
  • pillow * test
  • pytest * test
  • ray * test
  • ruff * test
  • tensordict * test
  • torch * test
  • torchvision * test
  • transformers * test
pyproject.toml pypi
requirements.txt pypi
  • accelerate *
  • codetiming *
  • datasets *
  • flash-attn >=2.4.3
  • liger-kernel *
  • mathruler *
  • numpy *
  • omegaconf *
  • pandas *
  • peft *
  • pillow *
  • pyarrow >=15.0.0
  • pylatexenc *
  • qwen-vl-utils *
  • ray *
  • tensordict *
  • torchdata *
  • transformers >=4.51.0,<4.53.0
  • vllm >=0.8.0
  • wandb *
setup.py pypi