feature: add more param

.gitattributes

@@ -1,14 +1,16 @@
-data/ltyai filter=lfs diff=lfs merge=lfs -text
-data/nanmei filter=lfs diff=lfs merge=lfs -text
-data/tianyi filter=lfs diff=lfs merge=lfs -text
-data/azusa filter=lfs diff=lfs merge=lfs -text
-data/encoder.pt filter=lfs diff=lfs merge=lfs -text
-data/g_hifigan.pt filter=lfs diff=lfs merge=lfs -text
-data/tianyi/record.wav filter=lfs diff=lfs merge=lfs -text
-data/tianyi/tianyi.pt filter=lfs diff=lfs merge=lfs -text
-data/azusa/azusa.pt filter=lfs diff=lfs merge=lfs -text
-data/azusa/record.wav filter=lfs diff=lfs merge=lfs -text
-data/ltyai/ltyai.pt filter=lfs diff=lfs merge=lfs -text
-data/ltyai/record.wav filter=lfs diff=lfs merge=lfs -text
-data/nanmei/nanmei.pt filter=lfs diff=lfs merge=lfs -text
-data/nanmei/record.wav filter=lfs diff=lfs merge=lfs -text
+*.bin.* filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tar.gz filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text

app.py

@@ -12,7 +12,7 @@ mockingbird = MockingBird()
 mockingbird_path = Path(os.path.dirname(__file__)) / "data"
 base_url = "https://al.smoe.top/d/Home/source/mockingbird/"
 
-for sy in ["encoder.pt", "g_hifigan.pt"]:
+for sy in ["encoder.pt", "g_hifigan.pt", "wavernn.pt"]:
     if not os.path.exists(os.path.join(mockingbird_path, sy)):
         torch.hub.download_url_to_file(f"{base_url}/{sy}", mockingbird_path / sy)
 
@@ -28,21 +28,28 @@ for model in ["azusa", "nanmei", "ltyai", "tianyi"]:
 mockingbird.load_model(
     Path(os.path.join(mockingbird_path, "encoder.pt")),
     Path(os.path.join(mockingbird_path, "g_hifigan.pt")),
-    # Path(os.path.join(mockingbird_path, "wavernn.pt"))
+    Path(os.path.join(mockingbird_path, "wavernn.pt")),
 )
 
 
-def inference(text: str, model_name: str):
+def inference(
+    text: str,
+    model_name: str,
+    vocoder_type: str = "HifiGan",
+    style_idx: int = 0,
+    min_stop_token: int = 9,
+    steps: int = 2000,
+):
     model_path = mockingbird_path / model_name
     mockingbird.set_synthesizer(Path(os.path.join(model_path, f"{model_name}.pt")))
     fd = NamedTemporaryFile(suffix=".wav", delete=False)
     record = mockingbird.synthesize(
         text=str(text),
         input_wav=model_path / "record.wav",
-        vocoder_type="HifiGan",
-        style_idx=0,
-        min_stop_token=9,
-        steps=2000,
+        vocoder_type=vocoder_type,
+        style_idx=style_idx,
+        min_stop_token=min_stop_token,
+        steps=steps,
     )
     with open(fd.name, "wb") as file:
         file.write(record.getvalue())
@@ -60,11 +67,20 @@ gr.Interface(
         gr.Radio(
             ["azusa", "nanmei", "ltyai", "tianyi"],
             label="model type",
+            value="azusa",
         ),
+        gr.Radio(
+            ["HifiGan", "WaveRNN"],
+            label="Vocoder type",
+            value="HifiGan",
+        ),
+        gr.Slider(minimum=-1, maximum=9, step=1, label="style idx", value=0),
+        gr.Slider(minimum=3, maximum=9, label="min stop token", value=9),
+        gr.Slider(minimum=200, maximum=2000, label="steps", value=2000),
     ],
     gr.Audio(type="filepath", label="Output"),
     title=title,
     description=description,
     article=article,
-    examples=[["阿梓不是你的电子播放器", "azusa"], ["不是", "nanmei"]],
+    examples=[["阿梓不是你的电子播放器", "azusa", "HifiGan", 0, 9, 2000], ["不是", "nanmei", "HifiGan", 0, 9, 2000]],
 ).launch()

data/wavernn.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d7a6861589e927e0fbdaa5849ca022258fe2b58a20cc7bfb8fb598ccf936169
+size 53845290

requirements.txt

@@ -10,5 +10,4 @@ webrtcvad
 Unidecode
 inflect
 loguru
-gradio
-tempfile
+gradio