During the dataset loading before training , it failed.
But when I put the loading script in a single file like test.py , the tar file can be successfully parsed and load:
ifile="data//train_v17/tr/data.list" # tar list including line like this /data/private/data/shard/train_v17/train_v17/tr/shards_000002632.tar
f=open(ifile,'r')
line = f.readline()
j=0
while line:
During the dataset loading before training , it failed. But when I put the loading script in a single file like test.py , the tar file can be successfully parsed and load:
`#!/bin/python import io import tarfile import logging import torchaudio
ifile="data//train_v17/tr/data.list" # tar list including line like this /data/private/data/shard/train_v17/train_v17/tr/shards_000002632.tar f=open(ifile,'r') line = f.readline() j=0 while line:
stream=tarfile.open("/data/private/data/shard/train_v17/train_v17/tr/shards_000002632.tar",mode="r|*")
print(j) ` it worked , and the torchaudio loaded successfully