Merge pull request #940 from c4fun/fastapi-cosyvoice2

Add a inference_instruct2 route to support and defaultly supports cosyvoice2 in fastapi server

runtime/python/fastapi/server.py

@@ -72,6 +72,14 @@ async def inference_instruct(tts_text: str = Form(), spk_id: str = Form(), instr
     model_output = cosyvoice.inference_instruct(tts_text, spk_id, instruct_text)
     return StreamingResponse(generate_data(model_output))
 
+@app.get("/inference_instruct2")
+@app.post("/inference_instruct2")
+async def inference_instruct2(tts_text: str = Form(), instruct_text: str = Form(), prompt_wav: UploadFile = File()):
+    prompt_speech_16k = load_wav(prompt_wav.file, 16000)
+    model_output = cosyvoice.inference_instruct2(tts_text, instruct_text, prompt_speech_16k)
+    return StreamingResponse(generate_data(model_output))
+
+
 
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
@@ -90,4 +98,4 @@ async def inference_instruct(tts_text: str = Form(), spk_id: str = Form(), instr
             cosyvoice = CosyVoice2(args.model_dir)
         except Exception:
             raise TypeError('no valid model_type!')
-    uvicorn.run(app, host="0.0.0.0", port=args.port)
+    uvicorn.run(app, host="0.0.0.0", port=args.port)