ttxskk
About train
Sorry, due to the license issue, the release of the training dataset *.npz is still in discussion, which is essential for training. But we are expected to release the training code and the preprocessed *.npz files or preprocessing method at the end of May.
Sorry, due to the license issue, the release of the training dataset *.npz is still in discussion, which is essential for training. But we are expected to release the training code and the preprocessed *.npz files or preprocessing method at the end of May.
Hi, any progress about training code or datasets?
Hi @imabackstabber ,
Thanks for your patience. We recently fixed some errors in the npz files mentioned in issue #10, resulting in improved performance on AGORA compared to our paper. However, we still need to conduct further validation. I will release all the files once the validation is complete. But if you need these npz files now, please feel free to let me know, and I can share a copy here.
Hi @imabackstabber ,
Thanks for your patience. We recently fixed some errors in the npz files mentioned in issue #10, resulting in improved performance on AGORA compared to our paper. However, we still need to conduct further validation. I will release all the files once the validation is complete. But if you need these npz files now, please feel free to let me know, and I can share a copy here.
Sure, I'll be happy if I could get one copy! Could you share it with a google drive link or something?
Thank you for your contribution. I would be incredibly grateful if you could kindly provide me with a copy as well.
Hi @imabackstabber @hengyuan-zhang-0, #15 has been merged into the main branch. If you encounter any issues, please feel free to reach out. If I don't respond in time, you can also email me at [email protected].
Hi @Maskros-new, Could you please check if agora_train_3840_w_occ_multi_2010.npz has been placed under data/preprocessed_npz/cache? From your log, I guess the npz file is not placed under the correct path.