Merge pull request #25 from stealthinu/onnx化

PyInstallerを試しやすいように整備しました

python/Pipfile

@@ -16,6 +16,7 @@ scikit-learn = "==1.0.2"
 sounddevice = "==0.4.4"
 SoundFile = "==0.10.3.post1"
 onnxruntime-directml = "1.13.1"
+pyinstaller = "*"
 
 [dev-packages]
 

python/install_pipenv.ps1

@@ -0,0 +1,10 @@
+pip install --upgrade pip
+pip install pipenv
+$pythonUserPath = python -m site --user-site
+$pythonUserPath = $pythonUserPath.Replace('site-packages', 'Scripts')
+$ENV:Path += ";" + $pythonUserPath
+$userPath = [System.Environment]::GetEnvironmentVariable("Path", "User")
+$userPath += ";" + $pythonUserPath
+[System.Environment]::SetEnvironmentVariable("Path", $userPath, "User")
+$ENV:PIPENV_VENV_IN_PROJECT = '.venv'
+[System.Environment]::SetEnvironmentVariable("PIPENV_VENV_IN_PROJECT", ".venv", "User")

python/makeexe.bat → python/makeexe.ps1

@@ -1 +1,2 @@
 pipenv run pyinstaller mmvc_client.py --add-binary "./.venv/Lib/site-packages/onnxruntime/capi/onnxruntime_providers_shared.dll;./onnxruntime/capi/" --add-binary "./.venv/Lib/site-packages/onnxruntime/capi/DirectML.dll;./onnxruntime/capi/" --collect-data librosa --onedir --clean -y
+pipenv run pyinstaller output_audio_device_list.py --onefile

python/onnx_bench.py

@@ -0,0 +1,78 @@
+import argparse
+import time
+import onnxruntime as ort
+import torch
+
+
+def get_args():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--input_onnx", required=True)
+    return parser.parse_args()
+
+
+def inspect_onnx(session):
+    print("inputs")
+    for i in session.get_inputs():
+        print("name:{}\tshape:{}\tdtype:{}".format(i.name, i.shape, i.type))
+    print("outputs")
+    for i in session.get_outputs():
+        print("name:{}\tshape:{}\tdtype:{}".format(i.name, i.shape, i.type))
+
+
+def benchmark(session):
+    dummy_specs = torch.rand(1, 257, 60)
+    dummy_lengths = torch.LongTensor([60])
+    dummy_sid_src = torch.LongTensor([0])
+    dummy_sid_tgt = torch.LongTensor([1])
+
+    use_time_list = []
+    for i in range(30):
+        start = time.time()
+        output = session.run(
+            ["audio"],
+            {
+                "specs": dummy_specs.numpy(),
+                "lengths": dummy_lengths.numpy(),
+                "sid_src": dummy_sid_src.numpy(),
+                "sid_tgt": dummy_sid_tgt.numpy()
+            }
+        )
+        use_time = time.time() - start
+        use_time_list.append(use_time)
+        #print("use time:{}".format(use_time))
+    use_time_list = use_time_list[5:]
+    mean_use_time = sum(use_time_list) / len(use_time_list)
+    print(f"mean_use_time:{mean_use_time}")
+
+
+def main(args):
+    ort_session_cpu = ort.InferenceSession(
+        args.input_onnx,
+        providers=["CPUExecutionProvider"])
+
+    ort_session_cuda = ort.InferenceSession(
+        args.input_onnx,
+        providers=["CUDAExecutionProvider"])
+
+    # DirectMLで動かすための設定
+    ort_options = ort.SessionOptions()
+    ort_options.execution_mode = ort.ExecutionMode.ORT_SEQUENTIAL
+    ort_options.enable_mem_pattern = False
+    ort_session_dml = ort.InferenceSession(
+        args.input_onnx,
+        sess_options=ort_options,
+        providers=["DmlExecutionProvider"])
+
+    print("vits onnx benchmark")
+    inspect_onnx(ort_session_cpu)
+    print("ONNX CPU")
+    benchmark(ort_session_cpu)
+    print("ONNX CUDA")
+    benchmark(ort_session_cuda)
+    print("ONNX DirectML")
+    benchmark(ort_session_dml)
+
+if __name__ == '__main__':
+    args = get_args()
+    print(args)
+    main(args)