suno-ai
No GPU being used
`/home/gucci/miniconda3/lib/python3.11/site-packages/torch/utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly. To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage() return self.fget.get(instance, owner)() /home/gucci/miniconda3/lib/python3.11/site-packages/torch/nn/utils/weight_norm.py:30: UserWarning: torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm. warnings.warn("torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.") /home/gucci/miniconda3/lib/python3.11/site-packages/transformers/models/encodec/modeling_encodec.py:120: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad(True), rather than torch.tensor(sourceTensor). self.register_buffer("padding_total", torch.tensor(kernel_size - stride, dtype=torch.int64), persistent=False)
NVIDIA-SMI 535.104.12 Driver Version: 535.104.12 CUDA Version: 12.2 pytorch 2.1.0 python 3.11 ubuntu 20.04.6
torch.cuda.is_available() return true
but no process found by nvidia-smi and the interface is very slow using more than 300 seconds to generate 4 seconds wav looks like no gpu acceleration
`+---------------------------------------------------------------------------------------+ | NVIDIA-SMI 535.104.12 Driver Version: 535.104.12 CUDA Version: 12.2 | |-----------------------------------------+----------------------+----------------------+ | GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+======================+======================| | 0 Tesla P4 On | 00000000:01:00.0 Off | 0 | | N/A 47C P8 7W / 75W | 0MiB / 7680MiB | 0% Default | | | | N/A | +-----------------------------------------+----------------------+----------------------+
+---------------------------------------------------------------------------------------+ | Processes: | | GPU GI CI PID Type Process name GPU Memory | | ID ID Usage | |=======================================================================================| | No running processes found | +---------------------------------------------------------------------------------------+ `
Make sure you have NVidia Cuda drivers Installed. Then install required pytorch version from here DIRECTLY into the Bark folder. Similar to this pip3 install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu124 --target c:\AI\Bark-Voice\ --upgrade
I had to explicitly tell the model to use the GPU with device_map and .to(model.device) for the processor
model_name = 'suno/bark'
wav_processor = AutoProcessor.from_pretrained(model_name)
wav_model = BarkModel.from_pretrained(model_name, device_map='cuda', torch_dtype=torch.float32)
inputs = wav_processor(sentence, voice_preset="v2/en_speaker6").to(wav_model.device)
audio_array = wav_model.generate(**inputs, min).cpu().numpy().squeeze().tolist()
