53 lines
1.8 KiB
Python
53 lines
1.8 KiB
Python
|
import torch
|
|||
|
|
|
|||
|
|
from infer.lib.infer_pack.models_onnx import SynthesizerTrnMsNSFsidM
|
|||
|
|
|
|||
|
|
|
|||
|
|
def export_onnx(ModelPath, ExportedPath):
|
|||
|
|
cpt = torch.load(ModelPath, map_location="cpu")
|
|||
|
|
cpt["config"][-3] = cpt["weight"]["emb_g.weight"].shape[0]
|
|||
|
|
vec_channels = 256 if cpt.get("version", "v1") == "v1" else 768
|
|||
|
|
|
|||
|
|
test_phone = torch.rand(1, 200, vec_channels) # hidden unit
|
|||
|
|
test_phone_lengths = torch.tensor([200]).long() # hidden unit 长度（貌似没啥用）
|
|||
|
|
test_pitch = torch.randint(size=(1, 200), low=5, high=255) # 基频（单位赫兹）
|
|||
|
|
test_pitchf = torch.rand(1, 200) # nsf基频
|
|||
|
|
test_ds = torch.LongTensor([0]) # 说话人ID
|
|||
|
|
test_rnd = torch.rand(1, 192, 200) # 噪声（加入随机因子）
|
|||
|
|
|
|||
|
|
device = "cpu" # 导出时设备（不影响使用模型）
|
|||
|
|
|
|||
|
|
net_g = SynthesizerTrnMsNSFsidM(
|
|||
|
|
*cpt["config"], is_half=False, version=cpt.get("version", "v1")
|
|||
|
|
) # fp32导出（C++要支持fp16必须手动将内存重新排列所以暂时不用fp16）
|
|||
|
|
net_g.load_state_dict(cpt["weight"], strict=False)
|
|||
|
|
input_names = ["phone", "phone_lengths", "pitch", "pitchf", "ds", "rnd"]
|
|||
|
|
output_names = [
|
|||
|
|
"audio",
|
|||
|
|
]
|
|||
|
|
# net_g.construct_spkmixmap(n_speaker) 多角色混合轨道导出
|
|||
|
|
torch.onnx.export(
|
|||
|
|
net_g,
|
|||
|
|
(
|
|||
|
|
test_phone.to(device),
|
|||
|
|
test_phone_lengths.to(device),
|
|||
|
|
test_pitch.to(device),
|
|||
|
|
test_pitchf.to(device),
|
|||
|
|
test_ds.to(device),
|
|||
|
|
test_rnd.to(device),
|
|||
|
|
),
|
|||
|
|
ExportedPath,
|
|||
|
|
dynamic_axes={
|
|||
|
|
"phone": [1],
|
|||
|
|
"pitch": [1],
|
|||
|
|
"pitchf": [1],
|
|||
|
|
"rnd": [2],
|
|||
|
|
},
|
|||
|
|
do_constant_folding=False,
|
|||
|
|
opset_version=13,
|
|||
|
|
verbose=False,
|
|||
|
|
input_names=input_names,
|
|||
|
|
output_names=output_names,
|
|||
|
|
)
|
|||
|
|
return "Finished"
|