search

llSourcell / How-to-Build-a-Biomedical-Startup

This is the code for "How to Build a Biomedical Startup" by Siraj Raval on Youtube
138 stars 77 forks source link

How to integrate Colab DotProduct model into Fultter app #1

Open pumpkinband opened 5 years ago

eagerashwani commented 5 years ago

Where is model? I can't find

pumpkinband commented 5 years ago

https://colab.research.google.com/drive/11hAr1qo7VCSmIjWREFwyTFblU2LVeh1R how can I integrate this model to flutter app @therichiekumar

harshit-saraswat commented 5 years ago

Anyone has any idea about how to do it? as the developer hasn't uploaded any pb or h5 file for the saved model. Please shed some light?

bnealey commented 5 years ago

In this video, around timestamp 45:20, converting it into a .tflite is one of the steps.

Check the dev's previous healthcare app to see a nicely integrated tensorflow lite.

harshit-saraswat commented 5 years ago

@bnealey actually in the original repository of Doc Product and also the colab file link given in the repository, there isn't any model file which we can convert into the .tflite format. Any ideas about it?

bnealey commented 5 years ago

@harshit-saraswat did you find the directory labelled GPT2Folder? The python code in cell 3 needs to be updated.

harshit-saraswat commented 5 years ago

@bnealey, I see. Plus they even removed the Embedding file link which they gave in their repository. We need an updated repository and model now.

lmcdo commented 5 years ago

@bnealey, could you say how to convert the model to hd5 ? :) No luck with RetrieveQA. Confused by the weights, embeddings, and nn's involved

Chrome3912 commented 5 years ago

try the api: https://pocket-derm.firebaseapp.com/

eagerashwani commented 5 years ago

@Chrome3912 please provide example.

Chrome3912 commented 5 years ago

@therichiekumar you have to purchase the api to get a key. But after that its a simple post request w/ json. Python example: 

import requests
r = requests.post(url='http://li1027-34.members.linode.com/predict/{APIKEY}', json = {"question":"my skin is itchy"} )

This should work once you get the API key image