Update app.py

app.py

@@ -2,10 +2,26 @@ import torch
 from peft import PeftModel
 import transformers
 import gradio as gr
+import os
+os.system('pip install voicefixer --upgrade')
+from voicefixer import VoiceFixer
+voicefixer = VoiceFixer()
+
+from TTS.api import TTS
+tts = TTS(model_name="tts_models/multilingual/multi-dataset/your_tts", progress_bar=False, gpu=True)
 
 import whisper
 model1 = whisper.load_model("small")
 
+import torchaudio
+from speechbrain.pretrained import SpectralMaskEnhancement
+
+enhance_model = SpectralMaskEnhancement.from_hparams(
+source="speechbrain/metricgan-plus-voicebank",
+savedir="pretrained_models/metricgan-plus-voicebank",
+run_opts={"device":"cuda"},
+)
+
 assert (
     "LlamaTokenizer" in transformers._import_structure["models.llama"]
 ), "LLaMA is now in HuggingFace's main branch.\nPlease reinstall it: pip uninstall transformers && pip install git+https://github.com/huggingface/transformers.git"
@@ -135,7 +151,22 @@ def evaluate(
         )
     s = generation_output.sequences[0]
     output = tokenizer.decode(s)
-    return output.split("### Response:")[1].strip()
+
+    tts.tts_to_file(output.split("### Response:")[1].strip(), 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, output.split("### Response:")[1].strip(), "enhanced.wav"]
 
 
 g = gr.Interface(
@@ -154,8 +185,13 @@ g = gr.Interface(
     outputs=[
         gr.inputs.Textbox(
             lines=5,
-            label="Output",
-        )
+            label="Speech to Text",
+        ),
+        gr.inputs.Textbox(
+            lines=5,
+            label="Alpaca Output",
+        ),
+        gr.Audio(label="Audio with Custom Voice"),
     ],
     title="🦙🌲 Alpaca-LoRA",
     description="Alpaca-LoRA is a 7B-parameter LLaMA model finetuned to follow instructions. It is trained on the [Stanford Alpaca](https://github.com/tatsu-lab/stanford_alpaca) dataset and makes use of the Huggingface LLaMA implementation. For more information, please visit [the project's website](https://github.com/tloen/alpaca-lora).",