Daisuke Niizumi

Hi @WhiteTeaDragon, thank you for your interest and for pointing this out. I think you are correct. I plan to fix this in near future, then let me ask you...

Hi @WhiteTeaDragon, Just FYI, the issue has been fixed in the repository with exactly the same change above. I have tested by pre-training on FSD50K and evaluating the pre-trained model...

@lucidrains Thanks for your suggestion for how to use spectrogram audio, one cent from me is about the number of the mel bins. It would be usually much more than...

@batrlatom Hey, excuse me to keep you waiting, but I finally made a Colab notebook for you: https://github.com/daisukelab/sound-clf-pytorch/blob/master/advanced/Perceiver_MelSpecAudio_Example_Colab.ipynb But I have to say this is not finished because it didn't...

@leo19941227 Hi, thank you for handling this issue! It is a lovely solution. (BTW, it is also what I have been doing with my linear evaluation.) I can also suggest...

@leo19941227 Thank you for your quick work! I just finished a simple test with BYOL-A changes at my end, so I will incorporate this API implementation and try it out....

@leo19941227 Hi, I was waiting for you actually :) Can you find my comment regarding the error above? It stops us from moving forward.

@leo19941227 Excuse me for the belated reply. (being busy for DCASE challenge ops these days) - An error happens at line 280 in downstream/runner.py if I add the `configure_stats` method...

@leo19941227 Dear Leo, shall we restart solving this ...? I understand that the current code is fairly different from what we were working on this. It could be better to...

@andi611 Hello Andy, thanks for your help! > One way is to implement this in you model's init. For example, the input dataset path is required in your model's init...