hiyouga/ChatGLM-Efficient-Tuning

Health v2 · maintenance only

Flags: archived

Adoption not part of the score

3713 stars · 462 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

A Python framework for parameter-efficient fine-tuning (LoRA, QLoRA, P-Tuning, RLHF) of the ChatGLM-6B and ChatGLM2-6B language models using Hugging Face PEFT and Transformers. It includes a Web UI for training and evaluation and an OpenAI-compatible demo API, but the project is no longer maintained in favor of LLaMA-Factory.

Use cases

• fine-tune ChatGLM-6B with LoRA on custom instruction data
• run QLoRA 4-bit training on a single consumer GPU
• train ChatGLM with RLHF
• merge LoRA weights back into the base model
• serve a fine-tuned ChatGLM model behind an OpenAI-compatible API
• fine-tune a Chinese chat model without full-model training

When to choose

• you specifically need to fine-tune ChatGLM-6B or ChatGLM2-6B and want a ready-made PEFT pipeline
• you want a historical reference implementation of LoRA/QLoRA/RLHF tuning for GLM models

When to avoid

• you want active maintenance or support for newer models - use LLaMA-Factory instead
• you need to fine-tune non-GLM language models
• you want production-grade training infrastructure

Facets

library · maturity abandoned

llm-training machine-learning sdk large-language-models machine-learning python chatglm peft lora qlora fine-tuning rlhf huggingface transformers superseded-by-llama-factory natural-language-processing gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="hiyouga/ChatGLM-Efficient-Tuning")