Skip to content

Make text-generation models that support images capable of running images #10

@inf3rnus

Description

@inf3rnus

@Swayam4414 Informed us today that they would like to use StanfordAIMI/CheXagent-8b but are unable to feed it images.

We explained that this is because if a model is marked as text-generation, it only accepts text. If a model belongs to the image-text-to-text category, it can be used to process images and text.

As a result, we'll be expanding the capabilities of text-generation models that support multi modal inputs, starting with StanfordAIMI/CheXagent-8b and then moving onto others that may meet the criteria.

Metadata

Metadata

Labels

enhancementNew feature or request

Type

No type
No fields configured for issues without a type.

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions