PeterPanUnderhill comments

Repositories
Issues
Comments

Results 5 comments of


                                            PeterPanUnderhill

Use BERT fine-tuned model for Tensorflow serving

> when i export my model as follows,how can i send batch requests to the server. > Thanks in advance for all the help. > > signature_def['serving_default']: > The given...

Key points to reproduce the result on your own dataset.

same question here

add @profile on method in self-defined package

same problem here

Generic version is still compiled with AVX2 support

Same problem here, would like to see updates, thx

I think there is a wrong in FocalLoss

If you set reduction as 'none', then it returns a tensor