Open OswaldoBornemann opened 1 month ago
It will be more robust if you split your manifest into parts and process each part separately. You can launch the script multiple times with GPU ID as an argument.
I see. Thank you.
So how to split the manifest like the cut
file did?
So suppose I have 4 million audio files, and I want to extract the features using
Encodec
. I found that the extracting process is kind of slow, whereComputing features in batches: 183301it [19:36, 176.40it/s]
. The extracting process only leverages single GPU. So I was wonder whether we can use multi-GPU to accelerate the extraction process in thecompute_and_store_features_batch
.