Jaime FH
Jaime FH
@ydennisy Ill check it out thanks!
@hirofumi0810 Hi again, So I tried to run the same steps as in the original post and now im stuck with the warprnnt make step. My output is: ``` git...
I found this PR on the repo with support for compute 30 https://github.com/HawkAaron/warp-transducer/pull/76, will give it a try and come back EDIT: Managed to compile it with the branch at...
@hirofumi0810 I managed to skip last problem skipping the data validation step (assumming al processing went right) and now I'm stuck at the LM training, it fails due to a...
@hirofumi0810 Hi! Thanks for the help. I tried that and now it fails in another step. It does starts the first minibatch though. ``` 0%| | 0/982390016 [00:00
@hirofumi0810 Hi Sorry Ive been out the last weeks, its a busy week for me this one but will try it on the weekend. Thanks
> @jiwidi how is it going here, still just a draft? 🐰 Hi! sorry for the delayed answer. There is some minor changes I still want to do. I will...
I would be interested in this model as well!
@zanshuxun Thanks! I saw it but I was looking for a torch implementation. Do you know if there are any plans to bring it to the torch version repo?
The "padding" is done while preprocessing the data. We explode the full ordered list of each user ratings into multiple subsequences. Let me illustrate: ``` # Example: Transform # +---+------------+...