Supported Tasks

The following Hugging Face tasks are natively supported on Azure Machine Learning and, so on, on Microsoft Foundry:

• embeddings (also known as feature-extraction)
• sentence-similarity
• text-ranking (also formerly known as sentence-ranking)
• automatic-speech-recognition
• text-to-speech (also formerly known as text-to-audio)
• speech-to-text
• translation
• text-translation
• question-answering
• text-classification
• fill-mask
• token-classification
• summarization (also known as text-summarization)
• text-generation (also know as either completions, chat-completion, text2text-generation or conversational)
• image-text-to-text (also know as chat-completion with vision-capabilities)
• image-classification
• image-segmentation
• object-detection
• text-to-image
• zero-shot-image-classification
• table-question-answering
• zero-shot-classification
• visual-question-answering
• mask-generation

With upcoming support for some of the following tasks:

• image-to-image
• text-to-image with LoRA
• image-feature-extraction
• image-to-text (also known as image-captioning)
• text-to-speech
• image-to-3d
• audio-to-audio (also known as speech-to-speech)
• text-to-video

More information about all the supported tasks at Hugging Face Tasks.

Update on GitHub