LlamaFactory Unified Fine-Tuning Framework Trends at 70,551 Stars — 100+ LLMs and VLMs With ACL 2024 Paper

GitHub·medium signalXBlueskyLinkedInCopy link

LlamaFactory, the unified efficient fine-tuning framework supporting 100+ LLMs and VLMs, is trending on GitHub at 70,551 stars. The framework supports fine-tuning models including DeepSeek, Gemma, Llama, Qwen, and others through a single interface with techniques like LoRA, QLoRA, and full fine-tuning. Originally presented at ACL 2024, it remains the go-to tool for developers who need to fine-tune open models without writing custom training loops — particularly relevant as the Llama 4, Qwen 3, and Gemma 4 model families have all launched in April 2026.