|
Download docs/development.md from Vedisasi/UltraThinking-LLM-Training: direct link, hf CLI and curl.
- Browser
- Download file 759 Bytes
-
https://huggingface.co/Vedisasi/UltraThinking-LLM-Training/resolve/main/docs/development.md
- Command line
-
hf download hf://Vedisasi/UltraThinking-LLM-Training/docs/development.md
-
curl -L -o development.md https://huggingface.co/Vedisasi/UltraThinking-LLM-Training/resolve/main/docs/development.md
759 Bytes
| # Development Guide | |
| ## Environment | |
| ```bash | |
| python -m venv .venv | |
| source .venv/bin/activate # Windows: .venv\Scripts\activate | |
| pip install -r requirements.txt | |
| pip install -e . | |
| ``` | |
| ## Code style & checks | |
| - Format: Black | |
| - Lint: flake8 | |
| - Types: mypy | |
| - Tests: pytest | |
| ```bash | |
| black src/ tests/ | |
| flake8 src/ tests/ | |
| mypy src/ | |
| pytest -q | |
| ``` | |
| ## Project layout | |
| - `train_ultrathink.py` β entry point for training | |
| - `src/models/` β model architecture and components | |
| - `src/data/` β datasets and loaders | |
| - `src/training/` β distributed, RLHF utilities | |
| - `src/evaluation/` β benchmarks | |
| ## Making changes | |
| 1. Create a branch | |
| 2. Add tests for new behavior | |
| 3. Update docs if user-facing changes | |
| 4. Run checks before opening a PR | |