1berry/wokada-voice-changer
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

WIP: DDSP_SVC support

Browse Source
This commit is contained in:
wataru 2023-03-29 23:11:03 +09:00
parent fa2365c95e
commit 06770746d9
2 changed files with 21 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
.gitignore vendored
View File

@ -23,6 +23,8 @@ server/model_hubert
server/model_so-vits-svc-40v2_tsukuyomi/ server/model_so-vits-svc-40v2_tsukuyomi/
server/model_so-vits-svc-40v2_amitaro/ server/model_so-vits-svc-40v2_amitaro/
server/model_so-vits-svc-40/ server/model_so-vits-svc-40/
server/model_so-vits-svc-40_mahiro/
server/model_so-vits-svc-40_amitaro/
model_DDSP-SVC/ model_DDSP-SVC/
server/model_sovits server/model_sovits
server/test server/test

23
server/voice_changer/DDSP_SVC/DDSP_SVC.py
View File

@ -86,6 +86,20 @@ class DDSP_SVC:
args.data.encoder_sample_rate, args.data.encoder_sample_rate,
args.data.encoder_hop_size, args.data.encoder_hop_size,
device="cpu") device="cpu")
# ort_options = onnxruntime.SessionOptions()
# ort_options.intra_op_num_threads = 8
# self.onnx_session = onnxruntime.InferenceSession(
# "model_DDSP-SVC/hubert4.0.onnx",
# providers=providers
# )
# inputs = self.onnx_session.get_inputs()
# outputs = self.onnx_session.get_outputs()
# for input in inputs:
# print("input::::", input)
# for output in outputs:
# print("output::::", output)
# f0dec # f0dec
self.f0_detector = vo.F0_Extractor( self.f0_detector = vo.F0_Extractor(
# "crepe", # "crepe",
@ -228,10 +242,11 @@ class DDSP_SVC:
mask = data[3] mask = data[3]
convertSize = data[4] convertSize = data[4]
vol = data[4] vol = data[5]
if vol < self.settings.silentThreshold: # if vol < self.settings.silentThreshold:
return np.zeros(convertSize).astype(np.int16) # print("threshold")
# return np.zeros(convertSize).astype(np.int16)
with torch.no_grad(): with torch.no_grad():
spk_id = torch.LongTensor(np.array([[int(1)]])) spk_id = torch.LongTensor(np.array([[int(1)]]))
@ -243,7 +258,7 @@ class DDSP_SVC:
self.args.data.sampling_rate, self.args.data.sampling_rate,
f0, f0,
self.args.data.block_size, self.args.data.block_size,
adaptive_key=float(0)) adaptive_key=float(3))
result = seg_output.squeeze().cpu().numpy() * 32768.0 result = seg_output.squeeze().cpu().numpy() * 32768.0
return np.array(result).astype(np.int16) return np.array(result).astype(np.int16)