Serveless memory problem when deploy Wav2Vec2 with custom inference code

[philschmid]

What is the model size of your custom model? and also how are you creating the model.tar.gz ? I might be possible the zip size or model size caused the issue.