search

0xSameer / ast

Code to train Automatic Speech-to-Text (AST) models
9 stars 4 forks source link

MFCC generation script missing. #1

Open imrnh opened 3 months ago

imrnh commented 3 months ago

!python beam.py -m experiments/es_en_20h -n 5 -k 5 -w 0.6 -s fisher_dev

Generated the following error:

Beam for: experiments/es_en_20h gpu: 0

Computing beam results 0%| | 0/3977 [00:00<?, ?it/s] Traceback (most recent call last): File "/content/beam.py", line 111, in for utt in nn.data_loader.get_batch(1, File "/content/dataloader.py", line 145, in get_batch batch_data["X"].append(self._load_speech(u, set_key, max_sp)) File "/content/dataloader.py", line 103, in _load_speech x_data = xp.load(utt_path)[:max_sp] File "/usr/local/lib/python3.10/dist-packages/cupy/_io/npz.py", line 64, in load obj = numpy.load(file, mmap_mode, allow_pickle) File "/usr/local/lib/python3.10/dist-packages/numpy/lib/npyio.py", line 427, in load fid = stack.enter_context(open(os_fspath(file), "rb")) FileNotFoundError: [Errno 2] No such file or directory: '../speech2text/mfcc_13dim/fisher_dev/20051009/20051009_182032_217_fsp-B-100.npy'

The documentation (.md and .pdf) doesn't seem consistent for creating .npy files. Most of the shell command either missing or path not included correctly. Any fix possible?

Using cuda12.2: nvcc --version output:

nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2023 NVIDIA Corporation Built on Tue_Aug_15_22:02:13_PDT_2023 Cuda compilation tools, release 12.2, V12.2.140 Build cuda_12.2.r12.2/compiler.33191640_0

imrnh commented 3 months ago

If possible, can you provide the following files ?

imrnh commented 3 months ago

@0xSameer is there any way I can get those .npy files needed to run the code?