Type: huggingface.image_text_to_text.ImageTextToText

Namespace: huggingface.image_text_to_text

Description

Answers questions or follows instructions given both an image and text. image, text, visual question answering, multimodal, VLM

Use cases:
- Visual question answering with free-form reasoning
- Zero-shot object localization or structure extraction via instructions
- OCR-free document understanding when combined with prompts
- Multi-turn, instruction-following conversations grounded in an image

Properties

Property Type Description Default
model hf.image_text_to_text The image-text-to-text model to use. {'type': 'hf.image_text_to_text', 'repo_id': '', 'path': None, 'variant': None, 'allow_patterns': None, 'ignore_patterns': None}
image image The image to analyze. {'type': 'image', 'uri': '', 'asset_id': None, 'data': None}
prompt str Instruction or question for the model about the image. Describe this image.
max_new_tokens int Maximum number of tokens to generate. 256

Outputs

Output Type Description
output str  

Metadata

Browse other nodes in the huggingface.image_text_to_text namespace.