Skip to content

Latest commit

 

History

History
44 lines (33 loc) · 1.03 KB

README.md

File metadata and controls

44 lines (33 loc) · 1.03 KB

entropix

Entropy Based Sampling and Parallel CoT Decoding

The goal is to replicate o1 style CoT with open source models

Current supported models: llama3.1+

Future supported models: DeepSeekV2+ Mistral Large (123B)

Getting Started

install poetry

curl -sSL https://install.python-poetry.org | python3 -

install rust to build tiktoken

curl --proto '=https' --tlsv1.3 https://sh.rustup.rs -sSf | sh

poetry install

poetry install

download weights (Base and Instruct)

poetry run python download_weights.py --model-id meta-llama/Llama-3.2-1B --out-dir weights/1B-Base
poetry run python download_weights.py --model-id meta-llama/Llama-3.2-1B-Instruct --out-dir weights/1B-Instruct

download tokenizer.model from huggingface (or wherever) into the entropix folder

run it

 PYTHONPATH=. poetry run python entropix/main.py

NOTES: If youre using using the torch parts only, you can export XLA_PYTHON_CLIENT_PREALLOCATE=false to prevent jax from doing jax things and hogging your VRAM