Assignment 9

For this assignment, I first picked the Card Art model in StyleGan and hosted it in the Glitch model.

I copied the URL and token to the environment in then template and it all worked out fine. The change was not really obvious. I also changed the limit to 100 to avoid reaching limit.

After successfully hosted runway on glitch, I tried the training in ML. I tried the text generation with Lord of the Rings. I set the training steps to default 1000.

The training took about 30 minutes. Initially, in the first 100 steps, only empty string "" were generated. I don't really know what's going on here. Later, text was generated.

I ran this model in the workspace and it was quite like style of Lord of the Rings

To employ the model on Glitch, I used the template runway-ml-gpt-api

However, it kept showing Daily Limit for GPU was reached even I have already changed the count in json file to 0 and daily limit to 1000. I was kind of confused about this part.

Page updated

Google Sites

Report abuse