GPU not being initialized (RVC)

[nanowell]

Latest binary version Nvidia GPU 3060 ti

C:\Tools\MMVCServerSIO>MMVCServerSIO.exe     -p 18888     --https false     --content_vec_500 pretrain/checkpoint_best_legacy_500.pt     --content_vec_500_onnx pretrain/checkpoint_best_legacy_500.onnx     --content_vec_500_onnx_on false     --hubert_base pretrain/hubert_base.pt     --hubert_base_jp pretrain/rinna_hubert_base_jp.pt     --hubert_soft pretrain/hubert/hubert-soft-0d54a1f4.pt     --nsf_hifigan pretrain/nsf_hifigan/model     --model_dir models     --samples samples.json
    Booting PHASE :__main__
    Voice Changerを起動しています。
+    Internal_Port:18888
    protocol: HTTP
    -- ---- --
    ブラウザで次のURLを開いてください.
    http://<IP>:<PORT>/
    多くの場合は次のいずれかのURLにアクセスすると起動します。
    http://localhost:18888/

    Booting PHASE :__main__
    Booting PHASE :MMVCServerSIO
VoiceChanger Initialized (GPU_NUM:1, mps_enabled:False)
 [INFO] [DSH]  voice-changer-native-client.exe
 [INFO] [DSH]  Creating WndMsg Listener Window
 [INFO] [DSH]  Get number of capabilities
 [INFO] [DSH]  Get stream caps: 0
 [INFO] [DSH]  Get stream caps: 1
 [INFO] [DSH]  Get stream caps: 2
 [INFO] [DSH]  Get stream caps: 3
 [INFO] [DSH]  Get stream caps: 4
 [INFO] [DSH]  Get stream caps: 5
 [INFO] [DSH]  Get stream caps: 6
 [INFO] [DSH]  Get stream caps: 7
 [INFO] [DSH]  Get stream caps: 8
 [INFO] [DSH]  Get stream caps: 9
 [INFO] [DSH]  Get stream caps: 10
 [INFO] [DSH]  Get stream caps: 11
 [INFO] [DSH]  Get stream caps: 12
 [INFO] [DSH]  Get stream caps: 13
 [INFO] [DSH]  Get stream caps: 14
 [INFO] [DSH]  Get stream caps: 15
 [INFO] [DSH]  Get stream caps: 16
 [INFO] [DSH]  Get stream caps: 17
 [INFO] [DSH]  Get stream caps: 18
 [INFO] [DSH]  Get stream caps: 19
 [INFO] [DSH]  Get stream caps: 20
 [INFO] [DSH]  Get stream caps: 21
 [INFO] [DSH]  Get stream caps: 22
 [INFO] [DSH]  Get stream caps: 23
 [INFO] [DSH]  Get stream caps: 24
 [INFO] [DSH]  Get stream caps: 25
 [INFO] [DSH]  Get stream caps: 26
 [INFO] [DSH]  Get stream caps: 27
 [INFO] [DSH]  Get stream caps: 28
 [INFO] [DSH]  Get stream caps: 29
 [INFO] [DSH]  Get stream caps: 30
 [INFO] [DSH]  Get stream caps: 31
 [INFO] [DSH]  Get stream caps: 32
 [INFO] [DSH]  Destroying parent object
 [INFO] [DSH]  Destroying WndMsg Listener Window
 [INFO] [DSH]  Destroyed window
 [INFO] [DSH]  Unregistered window class
[2023-05-18 19:24:48] connet sid : 93GeriGUt4qVrYZvAAAB
[2023-05-18 19:24:48] connet sid : oiumqT8pfUdW6qaDAAAD
2023-05-18 19:24:49 | INFO | faiss.loader | Loading faiss with AVX2 support.
2023-05-18 19:24:49 | INFO | faiss.loader | Could not load library with AVX2 support due to:
ModuleNotFoundError("No module named 'faiss.swigfaiss_avx2'")
2023-05-18 19:24:49 | INFO | faiss.loader | Loading faiss.
2023-05-18 19:24:49 | INFO | faiss.loader | Successfully loaded faiss.
RVC initialization:  VoiceChangerParams(model_dir='models', samples='samples.json', content_vec_500='pretrain/checkpoint_best_legacy_500.pt', content_vec_500_onnx='pretrain/checkpoint_best_legacy_500.onnx', content_vec_500_onnx_on=0, hubert_base='pretrain/hubert_base.pt', hubert_base_jp='pretrain/rinna_hubert_base_jp.pt', hubert_soft='pretrain/hubert/hubert-soft-0d54a1f4.pt', nsf_hifigan='pretrain/nsf_hifigan/model')

[w-okada]

??? way did you think so?

[nanowell]

If gpu is being utilized it's not being utilized efficiently and most of the work is on cpu to process samples this creates an unnecessary output delay. Is it possible to configure your repo in order to utilize at least 50% of my gpu?

[SirBitesalot]

I thin I might encounter the same problem in #246

[KarinMiku]

i think i have the same issue, gpu is only being used like 5-6% at most (rtx 4070)

[w-okada]

Not all processes can be executed on the GPU. For instance, dio and harvesting only work on the CPU. Harvesting is a very computationally-intensive process. If you want to reduce the load on the CPU, you should use dio. However, dio does not produce results of the same quality as harvesting. In the next version, we will be able to use crepe, which can be used instead of harvesting and run on the GPU.

[nanowell]

Thank you. I am waiting for crepe to improve latency.

[nanowell]

.