audio base64

app.py

@@ -1,14 +1,13 @@
 import argparse
 import glob
-import os
 import os.path
-from sys import exit
-import shutil
+
+import PIL
+import PIL.ImageColor
 import gradio as gr
+import gradio_client.utils as gr_processing_utils
 import numpy as np
 import onnxruntime as rt
-import PIL
-import PIL.ImageColor
 import tqdm
 from huggingface_hub import hf_hub_download
 
@@ -18,6 +17,19 @@ from midi_tokenizer import MIDITokenizer
 
 in_space = os.getenv("SYSTEM") == "spaces"
 
+audio_postprocess_ori = gr.Audio.postprocess
+
+
+def audio_postprocess(self, y):
+    data = audio_postprocess_ori(self, y)
+    if data is None:
+        return None
+    return gr_processing_utils.encode_url_or_file_to_base64(data["name"])
+
+
+gr.Audio.postprocess = audio_postprocess
+
+
 def softmax(x, axis):
     x_max = np.amax(x, axis=axis, keepdims=True)
     exp_x_shifted = np.exp(x - x_max)