Daisuke Niizumi
Daisuke Niizumi
Hello @leo19941227, Thank you for your help. This is just a quick idea: It would be nice if the model could receive information about the downstream task in `model.__init__`. Or,...
@andi611 Hi Andy, thanks! You are correct if we want to switch file-based configuration. Though this already allows us to on-the-fly configuration in the command line, we would need to...
Dear @leo19941227, I have prepared a fix for this issue, then I'd like to make a pull request. However, I cannot make it to the main branch and am not...
Hi, I guess this will show you how. https://github.com/qiuqiangkong/panns_transfer_to_gtzan/blob/master/pytorch/models.py#L186-L188 The checkpoint file contains several contents. You would need 'model' only.
Let me reopen this. Please let me know if anything needs to be corrected.
Thanks for your question about Masked Modeling Duo. We plan to release the code for SUPERB in May or June if anything goes well with Interspeech. (We have one more...
@leo19941227 Hi, I made all the changes needed. Could you take a look again, please? Thanks!
No problem, I'm on vacation, too ;) (submission complete!)
Hello @kaiw7 and Yuan, I hate to cut in and share findings from my recent paper, but there's one thing I can share: the typical patch size of 16x16 might...
Hi @YuanGongND, Thank you for your valuable comment! I found that I totally missed that you already discussed that in Section 3.8 (Comparing Patch-based and Frame-based AST) in the SSAST...