JFan

It's not directly supported, but its possible by cloning the repo and linking to your google drive. Here's my example, though it's really messy: https://colab.research.google.com/drive/1qDS7ZvpJrAo940xVdQn_ZCEozDgiisFy

It doesn't look like you've downloaded the COCO train2014 folder, which contains the image files needed to train the model. Use "setup.sh" to download and unzip that. Use the argument...

That line of code `line 31` is for preprocessing data in optimize.py. It does not affect training, it only loads files into memory. `line 42` ``with tf.Graph().as_default(), tf.Session() as sess:``...

I don't think that's possible, since this repo doesn't have a way to resume training (yet) and from my experiments on a Xeon (6-core) @ 2.20GHz, it won't even make...

All the way at the end, it states that permission was denied. Perhaps try to specify the file in the `input` folder, and write a full path?

You could always use photoshop, but I don't think masking will be implemented any time soon

This is due to arguments being missing when you run the file. arguments are variables that are necessary to operation, and must be specified by the user. most arguments are...

Not everyone has the time (or the money to buy a GPU) in order to train a model, but if you would be willing to train one and share it...

Assuming that it's a Tesla K80 (hey lol it's same as the one I use) it'll take approx 7-8 hours on a 1920x1080x3 .jpeg style image on the ADAM optimizer....

Most definitely, @kaisark, If you can send me the video you've been using as your benchmark, I can send a version from my training results. The left is the style...