Description
Transformer-based models can be loaded, fine-tuned, and run for text, vision, audio, and multimodal workflows. Researchers and ML engineers can use pretrained models for inference, training, and evaluation. It may download models and process sensitive user or dataset content.