Upload folder using huggingface_hub

tts_gradio.py

@@ -10,7 +10,7 @@ import os
 import re
 import tempfile
 # import librosa
-# import numpy as np
+import numpy as np
 # import torch
 # from torch import no_grad, LongTensor
 # import commons
@@ -166,7 +166,8 @@ def openai(text, name):
     # exactly as it was in the original request.
     #data = '{\n    "model": "tts-1",\n    "input": "The quick brown fox jumped over the lazy dog.",\n    "voice": "alloy"\n  }'
     #response = requests.post('https://api.openai.com/v1/audio/speech', headers=headers, data=data)
-    return "Success", response
+    out_arr = np.frombuffer(response.content, dtype=np.uint8)
+    return "Success", (24000,out_arr)
 
 def elevenlabs(text,name):
     url = f"https://api.elevenlabs.io/v1/text-to-speech/{eleven_id_model_name_dict[name]}"
@@ -221,11 +222,11 @@ def microsoft(text, name, style="Neural"):
         data=data,
     )
     # breakpoint()
-    timestamp = int(time.time()*10000)
-    path = f'/tmp/output_{timestamp}.wav'
-    with open(path, 'wb') as f:
-        f.write(response.content)
-    return "Success", path
+    # timestamp = int(time.time()*10000)
+    # path = f'/tmp/output_{timestamp}.wav' # TODO: disk might full.
+    # with open(path, 'wb') as f:
+    #     f.write(response.content)
+    return "Success", response
     
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
@@ -271,6 +272,7 @@ if __name__ == '__main__':
     app.queue(max_size=10)
     app.launch(share=True)
     # _, audio = microsoft(all_example,microsoft_model_list[0])
+    # breakpoint()
     # print(audio)
     # with open("test99.mp3", "wb") as f:
     #     f.write(audio.content)