Merge pull request #213 from jiangyuxiaoxiao/master

update server_fastapi.py: 添加2.1模型推理支持。

get_emo.py

@@ -17,7 +17,7 @@ def get_emo(path):
     wav, sr = librosa.load(path, 16000)
     device = config.bert_gen_config.device
     return process_func(
-        np.expand_dims(wav, 0).astype(np.float),
+        np.expand_dims(wav, 0).astype(np.float64),
         sr,
         model,
         processor,

server_fastapi.py

@@ -5,11 +5,10 @@ import logging
 import gc
 import random
 
-from pydantic import BaseModel
 import gradio
 import numpy as np
 import utils
-from fastapi import FastAPI, Query, Request
+from fastapi import FastAPI, Query, Request, File, UploadFile, Form
 from fastapi.responses import Response, FileResponse
 from fastapi.staticfiles import StaticFiles
 from io import BytesIO
@@ -180,10 +179,7 @@ if __name__ == "__main__":
     async def index():
         return FileResponse("./Web/index.html")
 
-    class Text(BaseModel):
+    async def _voice(
-        text: str
-
-    def _voice(
         text: str,
         model_id: int,
         speaker_name: str,
@@ -195,7 +191,10 @@ if __name__ == "__main__":
         language: str,
         auto_translate: bool,
         auto_split: bool,
+        emotion: Optional[int] = None,
+        reference_audio=None,
     ) -> Union[Response, Dict[str, any]]:
+        """TTS实现函数"""
         # 检查模型是否存在
         if model_id not in loaded_models.models.keys():
             return {"status": 10, "detail": f"模型model_id={model_id}未加载"}
@@ -214,28 +213,12 @@ if __name__ == "__main__":
             language = loaded_models.models[model_id].language
         if auto_translate:
             text = trans.translate(Sentence=text, to_Language=language.lower())
-        if not auto_split:
+        if reference_audio is not None:
-            with torch.no_grad():
+            with BytesIO(await reference_audio.read()) as ref_audio:
-                audio = infer(
+                if not auto_split:
-                    text=text,
+                    with torch.no_grad():
-                    sdp_ratio=sdp_ratio,
+                        audio = infer(
-                    noise_scale=noise,
+                            text=text,
-                    noise_scale_w=noisew,
-                    length_scale=length,
-                    sid=speaker_name,
-                    language=language,
-                    hps=loaded_models.models[model_id].hps,
-                    net_g=loaded_models.models[model_id].net_g,
-                    device=loaded_models.models[model_id].device,
-                )
-        else:
-            texts = cut_sent(text)
-            audios = []
-            with torch.no_grad():
-                for t in texts:
-                    audios.append(
-                        infer(
-                            text=t,
                             sdp_ratio=sdp_ratio,
                             noise_scale=noise,
                             noise_scale_w=noisew,
@@ -245,11 +228,34 @@ if __name__ == "__main__":
                             hps=loaded_models.models[model_id].hps,
                             net_g=loaded_models.models[model_id].net_g,
                             device=loaded_models.models[model_id].device,
+                            emotion=emotion,
+                            reference_audio=ref_audio,
                         )
-                    )
+                        audio = gradio.processing_utils.convert_to_16_bit_wav(audio)
-                    audios.append(np.zeros(int(44100 * 0.2)))
+                else:
-                audio = np.concatenate(audios)
+                    texts = cut_sent(text)
-                audio = gradio.processing_utils.convert_to_16_bit_wav(audio)
+                    audios = []
+                    with torch.no_grad():
+                        for t in texts:
+                            audios.append(
+                                infer(
+                                    text=t,
+                                    sdp_ratio=sdp_ratio,
+                                    noise_scale=noise,
+                                    noise_scale_w=noisew,
+                                    length_scale=length,
+                                    sid=speaker_name,
+                                    language=language,
+                                    hps=loaded_models.models[model_id].hps,
+                                    net_g=loaded_models.models[model_id].net_g,
+                                    device=loaded_models.models[model_id].device,
+                                    emotion=emotion,
+                                    reference_audio=ref_audio,
+                                )
+                            )
+                            audios.append(np.zeros(int(44100 * 0.2)))
+                        audio = np.concatenate(audios)
+                        audio = gradio.processing_utils.convert_to_16_bit_wav(audio)
         with BytesIO() as wavContent:
             wavfile.write(
                 wavContent, loaded_models.models[model_id].hps.data.sampling_rate, audio
@@ -258,9 +264,9 @@ if __name__ == "__main__":
             return response
 
     @app.post("/voice")
-    def voice(
+    async def voice(
         request: Request,  # fastapi自动注入
-        text: Text,
+        text: str = Form(...),
         model_id: int = Query(..., description="模型ID"),  # 模型序号
         speaker_name: str = Query(
             None, description="说话人名"
@@ -273,13 +279,14 @@ if __name__ == "__main__":
         language: str = Query(None, description="语言"),  # 若不指定使用语言则使用默认值
         auto_translate: bool = Query(False, description="自动翻译"),
         auto_split: bool = Query(False, description="自动切分"),
+        emotion: Optional[int] = Query(None, description="emo"),
+        reference_audio: UploadFile = File(None),
     ):
-        """语音接口"""
+        """语音接口，若需要上传参考音频请仅使用post请求"""
-        text = text.text
         logger.info(
             f"{request.client.host}:{request.client.port}/voice  { unquote(str(request.query_params) )} text={text}"
         )
-        return _voice(
+        return await _voice(
             text=text,
             model_id=model_id,
             speaker_name=speaker_name,
@@ -291,10 +298,12 @@ if __name__ == "__main__":
             language=language,
             auto_translate=auto_translate,
             auto_split=auto_split,
+            emotion=emotion,
+            reference_audio=reference_audio,
         )
 
     @app.get("/voice")
-    def voice(
+    async def voice(
         request: Request,  # fastapi自动注入
         text: str = Query(..., description="输入文字"),
         model_id: int = Query(..., description="模型ID"),  # 模型序号
@@ -309,12 +318,13 @@ if __name__ == "__main__":
         language: str = Query(None, description="语言"),  # 若不指定使用语言则使用默认值
         auto_translate: bool = Query(False, description="自动翻译"),
         auto_split: bool = Query(False, description="自动切分"),
+        emotion: Optional[int] = Query(None, description="emo"),
     ):
         """语音接口"""
         logger.info(
             f"{request.client.host}:{request.client.port}/voice  { unquote(str(request.query_params) )}"
         )
-        return _voice(
+        return await _voice(
             text=text,
             model_id=model_id,
             speaker_name=speaker_name,
@@ -326,6 +336,7 @@ if __name__ == "__main__":
             language=language,
             auto_translate=auto_translate,
             auto_split=auto_split,
+            emotion=emotion,
         )
 
     @app.get("/models/info")