Update export_onnx.py

export_onnx.py

@@ -1,54 +1,54 @@
-from infer_pack.models_onnx import SynthesizerTrnMsNSFsidM
-import torch
-
-if __name__ == "__main__":
-    MoeVS = True  # 模型是否为MoeVoiceStudio（原MoeSS）使用
-
-    ModelPath = "Shiroha/shiroha.pth"  # 模型路径
-    ExportedPath = "model.onnx"  # 输出路径
-    hidden_channels = 256  # hidden_channels，为768Vec做准备
-    cpt = torch.load(ModelPath, map_location="cpu")
-    cpt["config"][-3] = cpt["weight"]["emb_g.weight"].shape[0]  # n_spk
-    print(*cpt["config"])
-
-    test_phone = torch.rand(1, 200, hidden_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
-    )  # 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=16,
-        verbose=False,
-        input_names=input_names,
-        output_names=output_names,
-    )
+from infer_pack.models_onnx import SynthesizerTrnMsNSFsidM
+import torch
+
+if __name__ == "__main__":
+    MoeVS = True  # Whether the model is used by MoeVoiceStudio (formerly MoeSS)
+
+    ModelPath = "Shiroha/shiroha.pth"  # model path
+    ExportedPath = "model.onnx"  # Output path
+    hidden_channels = 256  # hidden_channels, prepare for 768Vec
+    cpt = torch.load(ModelPath, map_location="cpu")
+    cpt["config"][-3] = cpt["weight"]["emb_g.weight"].shape[0]  # n_spk
+    print(*cpt["config"])
+
+    test_phone = torch.rand(1, 200, hidden_channels)  # hidden unit
+    test_phone_lengths = torch.tensor([200]).long()  # Hidden unit length (seemingly useless)
+    test_pitch = torch.randint(size=(1, 200), low=5, high=255)  # Fundamental frequency (in Hz)
+    test_pitchf = torch.rand(1, 200)  # NSF base frequency
+    test_ds = torch.LongTensor([0])  # Speaker ID
+    test_rnd = torch.rand(1, 192, 200)  # Noise (adding random factors)
+
+    device = "cpu"  # Device when exporting (does not affect the use of the model)
+
+    net_g = SynthesizerTrnMsNSFsidM(
+        *cpt["config"], is_half=False
+    )  # Device when exporting (does not affect the use of the model)
+    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) export of multiple role mixed tracks
+    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=16,
+        verbose=False,
+        input_names=input_names,
+        output_names=output_names,
+    )