search

RetroCirce / HTS-Audio-Transformer

The official code repo of "HTS-AT: A Hierarchical Token-Semantic Audio Transformer for Sound Classification and Detection"
https://arxiv.org/abs/2202.00874
MIT License
341 stars 62 forks source link

How can run this project with one GPU?! #32

Closed saeedmaroof closed 1 year ago

saeedmaroof commented 1 year ago

hi. I am trying to test this model on Google Speech Command but after testing progress bar complete, I get this error: Default process group has not been initialized, please make sure to call init_process_group

I google this error and found that this error happens because of 'SyncBatchNorm' in 1 GPU and I should replace them by normal ones. in your code i found 'sync_batchnorm' parameter in pl.Trainer() and set its value to 'False' and run test command again but it wasn't work.

could anyone please tell me how can i run this project on 1 GPU ?! thanks.

JonathanFL commented 1 year ago

Might be related to #21

RetroCirce commented 1 year ago

Hi,

Sorry for the late reply, I was extremely busy these days for different dues. Yes, you can refer to #21 . Also, you don't need to change the sync_batchnorm, just remain it True, it will automatically turn to false if it is not run on multi-gpu.