Pedro Gabriel Gengo Lourenço

Can you put some inputs and output examples?

Ok, I got your point. The issue for this behavior is because it is not the same if you set: `num_inference_steps = 2` and `strength = 1.`, and `num_inference_steps =...

@tomaarsen should we include the softmax operation inside the ONNX model? Or returning the logits is the desired behavior?

Don't run the CI yet, Im still making some tests

@tomaarsen the issue with the tests is a little bit harder: for models that uses `token_type_ids` the tests run like a charm. However, for models that don't use `token_type_ids` the...

I will try more things today, but by now, I'm showing a message when the model doesn't use the `token_type_ids`, just to validate if the whole flow is working as...

@tomaarsen I found a way to keep the same interface we are using today and force the `token_type_ids` to appear. It can't be an optional argument, because one time you...

@tomaarsen can we merge this? If yes, I will solve the conflict

I will run the script with this branch, but on my tests I was seeing discrepancy between the results too, and after the fix it worked and returned the same...

Hey @tomaarsen, can I work on this?