Fine-Tuning & Training

Libraries and platforms for fine-tuning, aligning and training open models on your own data, from LoRA on one GPU to multi-node runs.

Unsloth
Fine-tune and run open LLMs faster with much less memory, on a single GPU.
LLaMA-Factory
Unified framework and web UI for fine-tuning over a hundred open LLMs and VLMs.
TRL
Hugging Face library for post-training models with SFT, DPO, GRPO and other RL methods.
Axolotl
Config-driven toolkit for fine-tuning open LLMs with LoRA, full fine-tunes and multi-GPU training.