henrywoo/chatllama

Health v2 · maintenance only

Flags: no_license

Adoption not part of the score

1201 stars · 129 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

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

library · maturity maintenance

llm-training machine-learning deep-learning large-language-models machine-learning deep-learning artificial-intelligence python rlhf llama chatgpt fine-tuning deepspeed single-gpu gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="henrywoo/chatllama")