sarthakyadav / leaf-pytorch Goto Github PK

where is the data preparation?

Hi @SarthakYadav

For Speech Command Classification, please let me know which dataset you used.
I can't find csv file in speech_commands v0.02.tar.gz.

Thansk.

Hi @SarthakYadav

thanks for the great code! I think it would be easier to use it if you would turn it into a pip-installable library. I could create a PR to do this if you are interested..

I am reading some paper saying ,before changing the signal into the feature images,they were using window function.I am wondering is it also in this code repository . I searched the raw_dataset.py. raw_transforms.py and train.py, but i coudn't find where to change signal.

btw,in the raw_transforms.py ,class Compose ,is it 's input is a image?still the same question,where you change the signal express

Hello @SarthakYadav . Thanks for your implementation. I am presently trying to apply LEAF for the task of emotion recognition on the IEMOCAP dataset. Have you tried that, or do you see any roadblocks if I use this implementation?

thanks for your code! help us a lot!
when i run train，i got this Error:

File "train.py", line 334, in
acc = train(ARGS)
File "train.py", line 149, in train
model = Classifier(cfg).to(device)
File "/home/xxx/Ben/code/leaf-pytorch/models/classifier.py", line 11, in init
self.features = get_frontend(cfg)
File "/home/xxx/Ben/code/leaf-pytorch/leaf_pytorch/frontend_helper.py", line 25, in get_frontend
fe = Leaf(use_legacy_complex=use_legacy_complex, initializer=initializer)
File "/home/xxx/Ben/code/leaf-pytorch/leaf_pytorch/frontend.py", line 56, in init
use_legacy_complex=use_legacy_complex
File "/home/xxx/Ben/code/leaf-pytorch/leaf_pytorch/convolution.py", line 45, in init
init_weights = initializer((self._filters, 2))
File "/home/xxx/Ben/code/leaf-pytorch/leaf_pytorch/initializers.py", line 18, in call
return gabor_filters.gabor_params_from_mels()
File "/home/xxx/Ben/code/leaf-pytorch/leaf_pytorch/filters.py", line 30, in gabor_params_from_mels
sqrt_filters = torch.sqrt(self.mel_filters())
File "/home/xxx/Ben/code/leaf-pytorch/leaf_pytorch/filters.py", line 49, in mel_filters
mel_filters = torchaudio.transforms.create_fb_matrix(
AttributeError: module 'torchaudio.transforms' has no attribute 'create_fb_matrix'

i found that maybe "create_fb_matrix" was deprecated （i'm using 0.11）. https://github.com/pytorch/audio/issues/1669。

how can i fix it. or can i using others replace it?

of cause i will go back to the past version and try.