search

w-okada / voice-changer

リアルタイムボイスチェンジャー Realtime Voice Changer
Other
16.6k stars 1.81k forks source link

GPU not being used #708

Closed resolver101757 closed 1 year ago

resolver101757 commented 1 year ago

Issue Type

Bug Report

vc client version number

MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.11

OS

windows 10

GPU

Quadro T2000

Clear setting

yes

Sample model

no

Input chunk num

yes

Wait for a while

The GUI successfully launched.

read tutorial

yes

Extract files to a new folder.

yes

Voice Changer type

RVC

Model type

ONNX

Situation

when running voice changing, I can select my graphics card on the main website but when i choose a model it doesn't select my NVidia graphics card for processing. This results in choppy voice when testing. I've also installed nzxt and it shows not activity on my graphics card.

image

I've looked at the terminal window and the suspicious lines are below and it says to run in verbose mode to more information.

2023-08-13 21:37:21.0754442 [W:onnxruntime:, session_state.cc:1030 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Some nodes were not assigned to the preferred execution providers which may or may not have an negative impact on performance. e.g. ORT explicitly assigns shape related ops to CPU to improve perf.

This line says to run it in verbose mode but not sure how to do this ?

2023-08-13 21:37:21.9033795 [W:onnxruntime:, session_state.cc:1032 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Rerunning with verbose output on a non-minimal build will show node assignments.

application window capture

No response

logs on terminal

C:\temp\MMVCServerSIO later version>MMVCServerSIO.exe -p 18888 --https false --content_vec_500 pretrain/checkpoint_best_legacy_500.pt --content_vec_500_onnx pretrain/content_vec_500.onnx --content_vec_500_onnx_on true --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 --crepe_onnx_full pretrain/crepe_onnx_full.onnx --crepe_onnx_tiny pretrain/crepe_onnx_tiny.onnx --rmvpe pretrain/rmvpe.pt --model_dir model_dir --samples samples.json Booting PHASE :main PYTHON:3.10.11 (tags/v3.10.11:7d4cc5a, Apr 5 2023, 00:38:17) [MSC v.1929 64 bit (AMD64)] Activating the Voice Changer. [Voice Changer] download sample catalog. samples_0003_t2.json [Voice Changer] download sample catalog. samples_0003_o2.json [Voice Changer] download sample catalog. samples_0003_d2.json [Voice Changer] model_dir is already exists. skip download samples. Internal_Port:18888 protocol: HTTP

Please open the following URL in your browser.
http://<IP>:<PORT>/
In many cases, it will launch when you access any of the following URLs.
http://127.0.0.1:18888/

[VCClient] Access http://127.0.0.1:18888/ [VCClient] wait web server...0 http://127.0.0.1:18888/ Booting PHASE :main Booting PHASE :MMVCServerSIO [Voice Changer] model slot is changed -1 -> 3 ................RVC [Voice Changer] [RVCr2] Creating instance VoiceChangerV2 Initialized (GPU_NUM(cuda):1, mps_enabled:False, onnx_device:GPU) [Voice Changer][RVC]: update_settings gpu:0 [Voice Changer][RVCr2] Initializing... 2023-08-13 21:36:46.3871647 [W:onnxruntime:, session_state.cc:1030 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Some nodes were not assigned to the preferred execution providers which may or may not have an negative impact on performance. e.g. ORT explicitly assigns shape related ops to CPU to improve perf. 2023-08-13 21:36:46.3962699 [W:onnxruntime:, session_state.cc:1032 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Rerunning with verbose output on a non-minimal build will show node assignments. [Voice Changer] generate new embedder. (no embedder) 2023-08-13 21:36:48.3854977 [W:onnxruntime:, session_state.cc:1030 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Some nodes were not assigned to the preferred execution providers which may or may not have an negative impact on performance. e.g. ORT explicitly assigns shape related ops to CPU to improve perf. 2023-08-13 21:36:48.3925917 [W:onnxruntime:, session_state.cc:1032 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Rerunning with verbose output on a non-minimal build will show node assignments. [Voice Changer] Loading index... Try loading... model_dir\3\added_IVF2736_Flat_nprobe_1_v2.index.bin GENERATE INFERENCER<voice_changer.RVC.inferencer.OnnxRVCInferencer.OnnxRVCInferencer object at 0x000002C66D7E3B80> GENERATE EMBEDDER<voice_changer.RVC.embedder.OnnxContentvec.OnnxContentvec object at 0x000002C66D7E3BE0> GENERATE PITCH EXTRACTOR<voice_changer.RVC.pitchExtractor.HarvestPitchExtractor.HarvestPitchExtractor object at 0x000002C66D7E3C70> [Voice Changer] [RVC] Initializing... done [Voice Changer][RVC]: update_settings f0Detector:dio [Voice Changer][RVC]: update_settings silentThreshold:0 [Voice Changer][RVC]: update_settings serverReadChunkSize:384 [Voice Changer][RVC]: update_settings extraConvertSize:4096 [Voice Changer][RVC]: update_settings modelSlotIndex:1691784476003 [VCClient] wait web server... done 200 [2023-08-13 21:37:15] connet sid : a4hqKBVxj19mQ1GMAAAC [2023-08-13 21:37:15] connet sid : L--pGRcxKCmEcL5NAAAD [Voice Changer] update configuration: modelSlotIndex 1691959040003 [Voice Changer] model slot is changed 3 -> 3 ................RVC [Voice Changer] [RVCr2] Creating instance VoiceChangerV2 Initialized (GPU_NUM(cuda):1, mps_enabled:False, onnx_device:GPU) Pipeline has been deleted [Voice Changer][RVC]: update_settings gpu:0 [Voice Changer][RVCr2] Initializing... 2023-08-13 21:37:21.0754442 [W:onnxruntime:, session_state.cc:1030 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Some nodes were not assigned to the preferred execution providers which may or may not have an negative impact on performance. e.g. ORT explicitly assigns shape related ops to CPU to improve perf. 2023-08-13 21:37:21.0873682 [W:onnxruntime:, session_state.cc:1032 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Rerunning with verbose output on a non-minimal build will show node assignments. [Voice Changer] generate new embedder. (anyway) 2023-08-13 21:37:21.8933872 [W:onnxruntime:, session_state.cc:1030 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Some nodes were not assigned to the preferred execution providers which may or may not have an negative impact on performance. e.g. ORT explicitly assigns shape related ops to CPU to improve perf. 2023-08-13 21:37:21.9033795 [W:onnxruntime:, session_state.cc:1032 onnxruntime::VerifyEachNodeIsAssignedToAnEp] Rerunning with verbose output on a non-minimal build will show node assignments. [Voice Changer] Loading index... Try loading... model_dir\3\added_IVF2736_Flat_nprobe_1_v2.index.bin GENERATE INFERENCER<voice_changer.RVC.inferencer.OnnxRVCInferencer.OnnxRVCInferencer object at 0x000002C66D874460> GENERATE EMBEDDER<voice_changer.RVC.embedder.OnnxContentvec.OnnxContentvec object at 0x000002C66D7E3B80> GENERATE PITCH EXTRACTOR<voice_changer.RVC.pitchExtractor.HarvestPitchExtractor.HarvestPitchExtractor object at 0x000002C66D875630> [Voice Changer] [RVC] Initializing... done [Voice Changer][RVC]: update_settings f0Detector:dio [Voice Changer][RVC]: update_settings silentThreshold:0 [Voice Changer][RVC]: update_settings serverReadChunkSize:384 [Voice Changer][RVC]: update_settings extraConvertSize:4096 [Voice Changer][RVC]: update_settings modelSlotIndex:1691959040003 Generated Strengths: for prev:(2048,), for cur:(2048,) [Voice Changer] warming up... generating sola buffer. [Voice Changer] update configuration: f0Detector harvest [Voice Changer][RVC]: update_settings f0Detector:harvest [Voice Changer] update configuration: f0Detector crepe [Voice Changer][RVC]: update_settings f0Detector:crepe [2023-08-13 21:38:53] connet sid : q3vgfc_wkGySwL0pAAAF

w-okada commented 1 year ago

use dio, increase chunk, decrease extra

resolver101757 commented 1 year ago

Used DIO, increased chunk to 256 and extra cant be decreased anymore, its on 4096. Still same issue, its not using my nvidia graphics card. Any ideas on why its not selecting my graphics card?

Thanks

w-okada commented 1 year ago

It doesn't seem like there are any abnormalities showing up in the logs. Why do you think the GPU isn't being used? Have you checked the performance monitor? If possible, please provide screenshots of the performance monitor before and after implementation.

resolver101757 commented 1 year ago

Close off as complete, all working now. the app (nzxt) that I used to check the GPU said it was running at zero but I wanted something to check each core when pasting the issue details here and found nvidia app control panel utilization and it said it was running at 60%. I found another issue that was affecting performance of software, correctly that and all working as expected