README.md

metadata

tags:
  - image-to-text
  - image-captioning
  - endpoints-template
license: apache-2.0

Fork of salesforce/BLIP for a image-captioning task on 🤗Inference endpoint.

This repo uses a custom handler that allows an Inference Endpoint to accept an array of URLs to be used by the BLIP model.

This repo also expands on florentgbelidji/blip_captioning

Expected Payload

{
  "inputs": [
    "https://url.to/image_1.jpg",
    "https://url.to/image_2.jpg",
  ]
}

Response Payload

{
  "captions": [
    "a caption for the first image", 
    "a caption for the second image"
  ]
}