Retrieval-based-Voice-Conversion-WebUI icon indicating copy to clipboard operation
Retrieval-based-Voice-Conversion-WebUI copied to clipboard

Not enough memory: you tried to allocate 20971520 bytes

Open Yo295 opened this issue 2 months ago • 1 comments

Sometimes when i try to train a model, it give me this error: not enough memory: you tried to allocate 20971520 bytes I am using an Intel i5 8400 CPU with 16gb of ram, and an AMD RX 6400 GPU (it doesn't get recognised/it's not compatible so it's not getting used) I do know these specs aren't the best for training, but i've already made a few models with it so it should work. Usually I just try again or restart the computer and it works but it's really persistent on a specific model i'm trying to continue training on. Anyone know why that happens and how I can fix it ? Thanks.

Yo295 avatar Oct 19 '25 16:10 Yo295

I have an Nvidia GPU, so I don't know exactly how to solve your problem, but I bet it is a similar process. The problem I had was that torch was utilizing my CPU instead of my GPU.

This fixed it, but this assumes you have CUDA cores, which is an Nvidia thing: pip install torch torchvision torchaudio --upgrade --index-url https://download.pytorch.org/whl/cu118

Here is something that may work for you, but it looks like it only works on Linux. Worth trying anyway though: pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm5.4.2

InTheProcess avatar Nov 11 '25 18:11 InTheProcess