Spaces:
Sleeping
Trying to reproduce something similar with Pytorch
Hello @fantaxy !
For one of my courses, I am working on a final project for which I decided to make a short story generator using Pytorch.
While doing research in this regard, I found you on Huggingface and this Fantasy-Novel Generator. I think it's absolutely fascinating!
I've never built something like this from scratch but I have experience with pytorch and various packages in python.
I want to build a similar generator and I was hoping that you could point me in the right direction since your generator works so well!
Thank you for your time :D
Hello @fantaxy !
For one of my courses, I am working on a final project for which I decided to make a short story generator using Pytorch.
While doing research in this regard, I found you on Huggingface and this Fantasy-Novel Generator. I think it's absolutely fascinating!I've never built something like this from scratch but I have experience with pytorch and various packages in python.
I want to build a similar generator and I was hoping that you could point me in the right direction since your generator works so well!Thank you for your time :D
Nice to meet you. If there's anything I can help you with, please let me know. I'll be happy to help.
Hi again @fantaxy ! :D
I have finished building everything and am currently in the process of training my model.
I am currently using 53 books as the basis of my database, all from gutenberg.org.
I cleaned them all up, tokenizing everything and created all the sequences and vocabulary. I ended up with 5,052,499 sequences and targets and 60245 words in vocabulary.
I custom-made my own dataset so I can get experience working with different packages and learning how to interact with websites using beautifulsoup4.
From your experience building the Fantasy-Novel generator, do you think this is enough or would you recommend that I get a larger dataset?
And after this, I want to also make a image generator like you did! Your work really inspired me!
P.S. I accidentally closed the discussion and then immediately repoened it lol
Hi again @fantaxy ! :D
I have finished building everything and am currently in the process of training my model.
I am currently using 53 books as the basis of my database, all from gutenberg.org.
I cleaned them all up, tokenizing everything and created all the sequences and vocabulary. I ended up with 5,052,499 sequences and targets and 60245 words in vocabulary.I custom-made my own dataset so I can get experience working with different packages and learning how to interact with websites using beautifulsoup4.
From your experience building the Fantasy-Novel generator, do you think this is enough or would you recommend that I get a larger dataset?And after this, I want to also make a image generator like you did! Your work really inspired me!
P.S. I accidentally closed the discussion and then immediately repoened it lol
Wow! You've done a great job with this excellent work. I think that should be sufficient. I'd love to see the results of what you've implemented. If you're able to share your dataset, I'd be happy to help with your work by utilizing it. Well done!