(wip)debug

tts.py

@@ -2,6 +2,7 @@ import os
 from dotenv import load_dotenv
 import random
 from gradio_client import Client, handle_file,file
+from huggingface_hub.constants import HF_TOKEN_PATH
 
 load_dotenv()
 
@@ -9,6 +10,8 @@ ZEROGPU_TOKENS = os.getenv("ZEROGPU_TOKENS", "").split(",")
 
 
 def get_zerogpu_token():
+    if not ZEROGPU_TOKENS or ZEROGPU_TOKENS == [""]:
+        return os.getenv("HF_TOKEN")
     return random.choice(ZEROGPU_TOKENS)
 
 
@@ -44,12 +47,11 @@ data = {"text": "string", "provider": "string", "model": "string"}
 
 def set_client_for_session(space:str, user_token=None):
     if user_token is None:
-        x_ip_token = get_zerogpu_token()
+
+        return Client(space, hf_token=get_zerogpu_token())
     else:
         x_ip_token = user_token
-
-    # The "gradio/text-to-image" space is a ZeroGPU space
-    return Client(space, headers={"X-IP-Token": x_ip_token})
+        return Client(space, headers={"X-IP-Token": x_ip_token})
 
 def predict_index_tts(text, user_token=None, reference_audio_path=None):
     client = set_client_for_session("kemuriririn/IndexTTS",user_token=user_token)