Update app.py

app.py

@@ -54,6 +54,7 @@ def generate_prompt(instruction, input=None):
 def evaluate(
 #    instruction,
     audio,
+    upload,
     input=None,
     token_count=200,
     temperature=1.0,
@@ -61,7 +62,7 @@ def evaluate(
     presencePenalty = 0.1,
     countPenalty = 0.1,
 ):
-
+    res = []
     # load audio and pad/trim it to fit 30 seconds
     audio = whisper.load_audio(audio)
     audio = whisper.pad_or_trim(audio)
@@ -116,13 +117,32 @@ def evaluate(
             out_last = i + 1
     gc.collect()
     torch.cuda.empty_cache()
-    yield out_str.strip()
+
+    res.append(out_str.strip())
+
+    tts.tts_to_file(res, speaker_wav = upload, language="en", file_path="output.wav")
+    
+    voicefixer.restore(input="output.wav", # input wav file path
+                    output="audio1.wav", # output wav file path
+                    cuda=True, # whether to use gpu acceleration
+                    mode = 0) # You can try out mode 0, 1, or 2 to find out the best result
+    
+    noisy = enhance_model.load_audio(
+    "audio1.wav"
+    ).unsqueeze(0)
+
+    enhanced = enhance_model.enhance_batch(noisy, lengths=torch.tensor([1.]))
+    torchaudio.save("enhanced.wav", enhanced.cpu(), 16000)
+
+    return [result.text, res, "enhanced.wav"]    
+#    yield out_str.strip()
 
 g = gr.Interface(
     fn=evaluate,
     inputs=[
 #        gr.components.Textbox(lines=2, label="Instruction", value="Tell me about ravens."),
         gr.Audio(source="microphone", label = "请开始对话吧！", type="filepath"),
+        gr.Audio(source="upload", label = "请上传您喜欢的声音(wav文件)", type="filepath"),
         gr.components.Textbox(lines=2, label="Input", placeholder="none"),
         gr.components.Slider(minimum=10, maximum=200, step=10, value=150), # token_count
         gr.components.Slider(minimum=0.2, maximum=2.0, step=0.1, value=1.0), # temperature
@@ -131,10 +151,15 @@ g = gr.Interface(
         gr.components.Slider(0.0, 1.0, step=0.1, value=0.4),  # countPenalty        
     ],
     outputs=[
+        gr.inputs.Textbox(
+            lines=1,
+            label="Speech to Text"",
+        ),
         gr.inputs.Textbox(
             lines=5,
-            label="Output",
-        )
+            label="Raven Output",
+        ),
+        gr.Audio(label="Audio with Custom Voice"),
     ],
     title="🥳💬💕 - TalktoAI，随时随地，谈天说地！",
     description="🤖 - 让有人文关怀的AI造福每一个人！AI向善，文明璀璨！TalktoAI - Enable the future！",