Cloud-CV / Fabrik

:factory: Collaboratively build, visualize, and design neural nets in browser
GNU General Public License v3.0
1.12k stars 235 forks source link

GCI: Added Tensorflow docs #452

Closed seventeen-iterator closed 5 years ago

seventeen-iterator commented 5 years ago

Added another documentation about usage of exported models in tensorflow.

thatbrguy commented 5 years ago

Hello @seventeen-iterator good work so far. However, most of the code had been covered in the previous PR right, so we would want something more novel.

Here's what we can try. Write code for sample usage for your own neural network model. You can create a simple CNN for instance using instructions in the same blog post.

Do the above in a separate python file. Update TODO in your current python file saying "Once graph export it is fixed, usage functionality from the new file can be inserted here".

seventeen-iterator commented 5 years ago

Hello, I added CNN that trained on the fashion MNIST dataset. I'm not sure how good is the model I created, but more than enough for a sample.

thatbrguy commented 5 years ago

Hello @seventeen-iterator I have made some suggestions on your submission. Check it out.

seventeen-iterator commented 5 years ago

Also added some progress bars to make it look nearly like keras

seventeen-iterator commented 5 years ago

Is this easy enough to understand now?

thatbrguy commented 5 years ago

Hello @seventeen-iterator, code looks good. CLI logger also looks good. I'm attaching a screenshot here for other mentors reference:

tf

Two things:

seventeen-iterator commented 5 years ago

@thatbrguy Hi! I wasn't launching it first, because I didn't have python2. Also tensorflow doesn't exist for Windows + Python2. But I managed to launch tensorflow_fashion_mnist.py on the virtual machine, and no errors caught. (I installed anaconda, tensorflow and tqdm first) The rest I'll satisfy later.

thatbrguy commented 5 years ago

@seventeen-iterator Oh you're using windows. Alright then. Address the other changes before your time runs out! (I believe you have 5 hours or so left). You can request for additional time if needed.

seventeen-iterator commented 5 years ago

Both lines executed without errors in python2.