eduardoleao052

8 ms·

1. 1.▲
   PyTorch in the Browser (JavaScript)(github.com)

   1 points by eduardoleao052 2y ago|1 comments
2. 2.▲
   by eduardoleao052 2y ago

   Train and deploy Neural Nets and Transformers in the browser with one simple tag!
3. 3.▲
   Deep Learning in JavaScript(eduardoleao052.github.io)

   4 points by eduardoleao052 2y ago|1 comments
4. 4.▲
   by eduardoleao052 2y ago

   Three months ago I made a post here with an early version of this project, and received some great feedback! One of the most asked features was GPU support, so I added it alongside many other improvements. Hope you like it!
5. 5.▲
   by eduardoleao052 2y ago

   Thanks for the feedback, glad you liked it! It does, and they work similarly to their PyTorch counterparts. The explanation is on the “Running it Yourself” part of the README.
6. 6.▲
   Deep Learning in JavaScript, with GPU Support(github.com)

   5 points by eduardoleao052 2y ago|3 comments
7. 7.▲
   by eduardoleao052 3y ago

   Thanks for the tip, will look into it! Yes, I think it would always be better to leave a "vanilla" option available.
8. 8.▲
   by eduardoleao052 3y ago

   That's an interesting idea as well, could definitely see that working in some use cases.
9. 9.▲
   by eduardoleao052 3y ago

   I’m planning on creating a small article explaining the syntax and the functionalities of the Deep Learning library. I think that could be a useful learning tool, do you think it would help?
10. 10.▲
    by eduardoleao052 3y ago

    I plan on adding Typescript support soon! Thanks for the feedback
11. 11.▲
    by eduardoleao052 3y ago

    I think that’s it. I’ll probably add that soon
12. 12.▲
    by eduardoleao052 3y ago

    It would be with some simple tweaks. For instance, the current block does not support Cross-Attention, just Self-Attention.
13. 13.▲
    by eduardoleao052 3y ago

    Thats true, it’s a limitation of working between these languages. I tried to mitigate it by using clear JSDoc, so that each variable pops up alongside an explanation when calling a function.
14. 14.▲
    by eduardoleao052 3y ago

    Thank you so much for the feedback! I’m open to pull requests, absolutely. And about the article, that’s a great tip, I’ll definitely do that as well.
15. 15.▲
    by eduardoleao052 3y ago

    Fair enough, I’m currently studying what would be the best option, but I think you are right!
16. 16.▲
    by eduardoleao052 3y ago

    Really cool implementation. Thanks for the feedback as well!
17. 17.▲
    by eduardoleao052 3y ago

    Cool! Will take a look
18. 18.▲
    by eduardoleao052 3y ago

    Thank you! That’s going to help out a lot!
19. 19.▲
    by eduardoleao052 3y ago

    Thank you for the feedback! If you have any questions or suggestions looking it over tomorrow, I’d love to hear them!
20. 20.▲
    by eduardoleao052 3y ago

    I had read about this as a possibility, but will definitely look more into it now, thank you for the tip!
21. 21.▲
    by eduardoleao052 3y ago

    Thanks for the tip! I’ll definitely take a look. Adding GPU support is my next step!
22. 22.▲
    by eduardoleao052 3y ago

    Adding GPU support soon is absolutely my goal in the future! I think a PyTorch-based JavaScript library could be useful, as PyTorch has been way more dominant than TensorFlow recently.
23. 23.▲
    by eduardoleao052 3y ago

    I centralized the entire backpropagation around the Operation objects. They store the data about the forward prop in the cache, and serve as the connections in the graphs between tensors. Each tensor has a “parent”, “child” and “operation”.
24. 24.▲
    by eduardoleao052 3y ago

    I can try implementing it in the future lol It would surely be a quality of life improvement. But with the current tools, I tried my best to make the syntax as similar as possible to PyTorch’s!
25. 25.▲
    by eduardoleao052 3y ago

    Could not agree more hahaha! I tried to work around it building methods like “torch.add(a, b)” for operator overloading and “torch.at(index)” for slicing. But it’s definitely not as seamless as these features you proposed.
26. 26.▲
    by eduardoleao052 3y ago

    Initially I'll stick to Javascript (plus some libraries to improve performance), but I'll likely look into that in the future.
27. 27.▲
    by eduardoleao052 3y ago

    Now that I've tried Javascript, I could branch out to other languages in the future. I chose Javascript initially for its popularity and easy use on web browsers.
28. 28.▲
    by eduardoleao052 3y ago

    That’s true. I myself learned ML in python first, and I’m just now trying to implement some stuff in JavaScript. I guess it could be useful to make web integration with neural networks easier in this case, or at least as an educational pac
29. 29.▲
    by eduardoleao052 3y ago

    I am currently studying Typescript and other JavaScript libraries to improve the library’s performance. So adding GPU support is in sight in the future.
30. 30.▲
    by eduardoleao052 3y ago

    You are absolutely right! Will solve that as soon as possible. Thanks for the feedback!