google

/

tapas-large-finetuned-wtq

like 92

Hello @muhammadbaasit , how did you convert the model in tf-lite? Was it done with optimum?

If that is the case, please open an issue there with a reproducible code example and the full error that you're seeing (or instructions showing the lack of correct model outputs).

Thanks!

Hello @lysandre ,
I haven't it done with optimum because it was showing error with table question answering model. I have converted model using TFLiteConverter. please refer the below link and check for both outputs.

link: https://colab.research.google.com/drive/1HwvjdgUBlKsTzcybhEg2xWVuFJx0WEpe?usp=sharing

@fxmarty would you know what may be happening here/how to better support with optimum?

Hello @fxmarty ,
as i see in the logs as of now optimum is not supported for the TFLite and ONNX. because of that i am trying to export it to Tf-Lite using TFLiteConverter