self-supervised-speech-recognition
self-supervised-speech-recognition copied to clipboard
mailong25
memory leak with stt.py
Hi, If you do an inference with a Transcriber object t.transcribe(..), after returning the result, it should release any resources related to that inference. But it stays on VRAM and after a few calls to t.transcribe(), I get CUDA related out of memory errors. Looking at nvidia-smi it shows the memory is still occupied even after the transcript has been returned.
It would be nice to have a long lived Transcriber object which can be reused avoiding the lengthy creation time. If you're busy please give me a hint on how it may be done so I can give it a shot and submit a PR. Thanks for your project.
hi, I have the same problem, have you slove it?
Hi, If you do an inference with a
Transcriberobjectt.transcribe(..), after returning the result, it should release any resources related to that inference. But it stays on VRAM and after a few calls tot.transcribe(), I get CUDA related out of memory errors. Looking atnvidia-smiit shows the memory is still occupied even after the transcript has been returned.It would be nice to have a long lived
Transcriberobject which can be reused avoiding the lengthy creation time. If you're busy please give me a hint on how it may be done so I can give it a shot and submit a PR. Thanks for your project.
HI I have the same problem, have you slove it?
