sanjuktasr

Is there any code in the repo that directly generates the features from audios in training input format?

What are the different format of data inputs for zipformer/pretrain in SSL recipe? How can we obtain them?

what are the changes to be made if we want too change the FRAME SIZE from 480 to 160?

1

by changing the the frame size from 480 to 160 and not changing anything else the code is not working. there is some memory issues? What are the respective changes...

Read seperate .jsonl.gz from fbank filter them and make a Cutset into single variable.

3

code : for items in mf2: test_item_cuts = librispeech.test_clean_cuts(items) # test_filtered_cuts = test_item_cuts.filter(remove_short_and_long_utt) if test_item_cuts is not None: print(items) test_filtered_cuts = test_item_cuts.filter(remove_short_and_long_utt) if count==0: #test_other_cuts = librispeech.test_other_cuts() test_clean_cuts = test_filtered_cuts...

I want to update the FRAME SIZE TO 160 anf PITCH frame size to 320?