Hey @citnaj, I finally got DeOldify (https://github.com/jantic/DeOldify ) working on @GoogleColab. So now anyone can run their own pictures using the pre-trained model on the Colab GPU.https://github.com/mc-robinson/random/blob/master/deoldify_attempt.ipynb …
-
Show this thread
-
Hmm, I run this snippet of code in colab and keep getting an error No module named 'torch.nn.utils.spectral_norm'
2 replies 0 retweets 0 likes -
Look in the comments for the updated repo ( twitter really needs that edit button)
1 reply 0 retweets 0 likes -
Yeap, that worked. Thanx :) Now we got another issue :) What's this server? Imagenet? Connecting to 94.100.113.181:80... failed: Connection timed out. Retrying.
1 reply 0 retweets 0 likes -
So http://www.image-net.org/api/text/imagenet.synset.geturls?wnid=n01440764 … is just a list of urls for different images. Very possible that wget gets held up at one of the urls, might be worth pausing and editing the list to skip tough images. Not sure what you mean by the "server" though?
2 replies 0 retweets 1 like -
Alright, yeap. Just checked out the list http://www.image-net.org/api/text/imagenet.synset.geturls?wnid=n01440764 … It's a bit weird. Most of images contains fish.
1 reply 0 retweets 0 likes -
LOL Yeah that was just a hack to work with the http://Fast.AI library so I could get it up and running yet not wait for all the images to load (so I took a subset). Again...didn't think this would get this much attention so soon. Sorry!
2 replies 0 retweets 3 likes -
Replying to @citnaj @alexpolymath and
If anybody wants to make that part of the code the opposite of dumb and send a pull request, please do! I tried cleaning up the code a bit after it went public but it still has embarrassing hacks littered in it.
1 reply 0 retweets 2 likes
...I'm still laughing about he fish thing 
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.