Fine-tuning AI tools

2× faster, lower-memory fine-tuning for open models.

Tune Gemini models on your data in Google Cloud.

TRL, PEFT and AutoTrain for post-training on the Hub.

Large-scale GPU cloud and AI Studio for training and inference.

Ray for distributed training, tuning and serving.

Config-driven fine-tuning for many model architectures.

GPU cloud for training and fine-tuning.

Supervised, DPO and reinforcement fine-tuning for GPT models.

Distill expensive prompts into cheap fine-tuned models; RL via ART.

Fine-tune and serve small models with LoRAX multi-adapter serving.