Text to Image metric incorrect evaluation

[shanmukh-sai2003]

❗BEFORE YOU BEGIN❗
Are you on discord? 🤗 We'd love to have you asking questions on discord instead: https://discord.com/invite/a3K9c8GRGt

Describe the bug
In Text to image metric while evaluating Semantic Consistency we should pass the input text prompt and actual output image to the model for evaluation. Instead, we are passing the input text prompt and image from the input if it contains any.

Expected behavior
We should pass the actual output image and input text prompt to the model for evaluation.

Screenshots

[kritinv]

Hey @shanmukh-sai2003, thanks for pointing it out. It should be fixed here: #1253

[shanmukh-sai2003]

thanks for the fix @kritinv. Can I know when I can expect this fix to be released

[penguine-ip]

@shanmukh-sai2003 I believe its already out, can you give it a try? :)