Hello.
I have finished training a BERT based model for a text classification task. The model first uses tensorflow hub BERT model to preprocess the input and it also has additional custom layers. I want to deploy it as a web application. How can I go about this? I tried using flask but the .h5 file access is not working.
How can I save this model to deploy for users?
Hi @Ananya ,
the .h5 file can be converted into a tensorflow.js web format using tensorflowjs_converter. Please have a look here:
There’re also examples below, how to load the new web model and run inference in the browser.
Let me know,
Dennis
thanks it worked but I have more problems now :')