CoT-Finetune-LLM

A simple script for distributed finetuning and teaching any typical Large Language model thinking using Chain of Thought learning methods

Based on

Original Script based on this disscussion [https://huggingface.co/datasets/KingNish/reasoning-base-20k/discussions/3#6708df6dcd35b971ea0fb6ff].

Dataset Used

Currently using [https://huggingface.co/datasets/KingNish/reasoning-base-20k] but [https://huggingface.co/datasets?search=cot] can be used to select other datasets as well.

Usage

Update the model name in the training.py file. Once update save he file and then launch the training using trco Save the script and

Huggingface Models

Few of these trained models are located at [https://huggingface.co/collections/ewre324/chain-of-thought-reasoning-677cd3d30a4d2007a88b159a]