Stanislau Hluboki

@ArneVogel any chance you can review it? :)

As mentioned in Matthew's blog, train.py takes about 6 hours using NVIDIA GTX 970.

I'm not sure whether I'm doing everything right, but after 150K batches I have very poor results, too. P.S. Running on Python 3.6 and TensorFlow 1.3. UPD. Decreased learning rate...

`sys.argv` holds command line arguments. You must pass an image as a first argument: ``` ./detect.py in.jpg ```

In my case, I am using email provider, so I had to do this in the middleware to avoid issues with the auth config validation: ```typescript import NextAuth from "next-auth"...