#load file with librosa
x, sr = librosa.load(file)
Tx, Wx, ssq_freqs, scales, *_ = ssq_cwt(x, nv=16, wavelet='morlet', astensor=True, cache_wavelet=True)
RuntimeError: CUDA out of memory. Tried to allocate 2.53 GiB (GPU 0; 7.79 GiB total capacity; 5.76 GiB already allocated; 947.44 MiB free; 5.76 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
I checked nvidia-smi and yes it starts taking memory until it reaches the mem limit, then it releases when it crashes
is there any trick to analyze such files without hitting the memory limits ? Thanks
trying to run with this file :
it's a 22MB wav file
first I load the file with librosa :
RuntimeError: CUDA out of memory. Tried to allocate 2.53 GiB (GPU 0; 7.79 GiB total capacity; 5.76 GiB already allocated; 947.44 MiB free; 5.76 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONFI checked nvidia-smi and yes it starts taking memory until it reaches the mem limit, then it releases when it crashes
is there any trick to analyze such files without hitting the memory limits ? Thanks