GPU Requirements
I'm facing GPU issue for this. What is the minimum GPU required for this?
I"m getting the following errors:
dist: False
24-04-20 16:58:10.817 - INFO: Random seed: 6473
24-04-20 16:58:52.494 - INFO: Number of training data elements: 118, iters: 1
24-04-20 16:58:52.494 - INFO: Total epochs needed: 450 for iters 450
F:\Content Creation\Voice Training\ai-voice-cloning-3.0\runtime\Lib\site-packages\transformers\configuration_utils.py:380: UserWarning: Passing gradient_checkpointing to a config initialization is deprecated and will be removed in v5 Transformers. Using model.gradient_checkpointing_enable() instead, or if you are using the Trainer API, pass gradient_checkpointing=True in your TrainingArguments.
warnings.warn(
24-04-20 16:59:50.891 - INFO: Loading model for [./models/tortoise/autoregressive.pth]
I'm also seeing this error in the logs:
[torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 12.00 MiB. GPU 0 has a total capacity of 2.00 GiB of which 0 bytes is free. Of the allocated memory 3.48 GiB is allocated by PyTorch, and 95.02 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management
im having the same issues
You need way more Dedicated GPU memory. Shared means its using the system RAM which is much slower. And I've seen mine go upto 14 GB of dedicated GPU memory being used so I would assume you need at least a 8 GB card and then have the other 6 GB spill over to system RAM. It will likely take 5-10 times longer though.