# henrywoo/chatllama

ChatLLaMA 📢 Open source implementation for LLaMA-based ChatGPT runnable in a single GPU. 15x faster training process than ChatGPT

Repository: https://github.com/henrywoo/chatllama
Canonical: https://ross.abutalabs.com/products/chatllama
Language: Python
License Family: other
Last push: 2025-01-18T15:53:00+00:00

## Health v2 (maintenance only)
Score: 22/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 2, release rhythm 8, longevity 91
- inputs: {"age_days": 1283, "days_push": 592, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1201, forks 129 (observed 2026-08-28T04:03:58.155409+00:00)

## What it is
An open-source Python library implementing RLHF-based fine-tuning of Meta's LLaMA models to build ChatGPT-style chat assistants, runnable on a single GPU with DeepSpeed ZERO support. It provides training pipelines for all LLaMA architectures (7B-65B) but does not include model weights.

## Use cases
- fine-tune llama with rlhf
- build an open-source chatgpt alternative
- train a chatbot on a single gpu
- reinforcement learning from human feedback for llama
- speed up llm fine-tuning with deepspeed
- create an instruction-tuned chat model

## When to choose
- you want an open-source RLHF training pipeline for LLaMA models
- you have limited GPU resources and need single-GPU fine-tuning
- you want DeepSpeed ZERO integration for faster training

## When to avoid
- you need ready-to-use model weights or inference without training
- you want a maintained project with active development and a license
- you need a production chatbot deployment rather than a training framework

## Facets
- artifact type: library
- maturity: maintenance
- function: llm-training, machine-learning, deep-learning
- domain: large-language-models, machine-learning, deep-learning, artificial-intelligence
- platform: python
- tags: rlhf, llama, chatgpt, fine-tuning, deepspeed, single-gpu, gpu, linux

## Member repositories
- henrywoo/chatllama (main) score 22

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:58.155409+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:20:41.759130+00:00, confidence not recorded.
  - readme: https://github.com/henrywoo/chatllama (fetched 2026-08-28T04:03:58.155409+00:00, sha b97915e4dfd3)
- Data as of 2026-08-30T08:39:29.467469+00:00.