(wip)debug

models.py

@@ -404,35 +404,6 @@ def insert_initial_models():
             is_open=True,
             model_url="https://github.com/Index-Research/index-tts",
         ),
-        # Model(
-        #     id="eleven-multilingual-v2",
-        #     name="Eleven Multilingual v2",
-        #     model_type=ModelType.TTS,
-        #     is_open=False,
-        #     model_url="https://elevenlabs.io/",
-        # ),
-        # Model(
-        #     id="eleven-turbo-v2.5",
-        #     name="Eleven Turbo v2.5",
-        #     model_type=ModelType.TTS,
-        #     is_open=False,
-        #     model_url="https://elevenlabs.io/",
-        # ),
-        # Model(
-        #     id="eleven-flash-v2.5",
-        #     name="Eleven Flash v2.5",
-        #     model_type=ModelType.TTS,
-        #     is_open=False,
-        #     model_url="https://elevenlabs.io/",
-        # ),
-        # Model(
-        #     id="cartesia-sonic-2",
-        #     name="Cartesia Sonic 2",
-        #     model_type=ModelType.TTS,
-        #     is_open=False,
-        #     is_active=False, # ran out of credits
-        #     model_url="https://cartesia.ai/",
-        # ),
         Model(
             id="spark-tts",
             name="Spark TTS",
@@ -446,30 +417,9 @@ def insert_initial_models():
             name="maskgct",
             model_type=ModelType.TTS,
             is_open=False,
-            is_active=True,
+            is_active=False,
             model_url="https://github.com/open-mmlab/Amphion/tree/main/models/tts/maskgct",
         ),
-        # Model(
-        #     id="playht-2.0",
-        #     name="PlayHT 2.0",
-        #     model_type=ModelType.TTS,
-        #     is_open=False,
-        #     model_url="https://play.ht/",
-        # ),
-        # Model(
-        #     id="styletts2",
-        #     name="StyleTTS 2",
-        #     model_type=ModelType.TTS,
-        #     is_open=True,
-        #     model_url="https://github.com/yl4579/StyleTTS2",
-        # ),
-        # Model(
-        #     id="kokoro-v1",
-        #     name="Kokoro v1.0",
-        #     model_type=ModelType.TTS,
-        #     is_open=True,
-        #     model_url="https://huggingface.co/hexgrad/Kokoro-82M",
-        # ),
         Model(
             id="cosyvoice-2.0",
             name="CosyVoice 2.0",
@@ -483,49 +433,7 @@ def insert_initial_models():
             model_type=ModelType.TTS,
             is_open=True,
             model_url="https://huggingface.co/spaces/lj1995/GPT-SoVITS-v2",
-        ),
-        # Model(
-        #     id="papla-p1",
-        #     name="Papla P1",
-        #     model_type=ModelType.TTS,
-        #     is_open=False,
-        #     model_url="https://papla.media/",
-        # ),
-        # Model(
-        #     id="hume-octave",
-        #     name="Hume Octave",
-        #     model_type=ModelType.TTS,
-        #     is_open=False,
-        #     model_url="https://hume.ai/",
-        # ),
-        # Model(
-        #     id="megatts3",
-        #     name="MegaTTS 3",
-        #     model_type=ModelType.TTS,
-        #     is_active=False,
-        #     is_open=True,
-        #     model_url="https://github.com/bytedance/MegaTTS3",
-        # ),
-        # Model(
-        #     id="minimax-02-hd",
-        #     name="MiniMax Speech-02-HD",
-        #     model_type=ModelType.TTS,
-        #     is_open=False,
-        #     model_url="http://minimax.io/",
-        # ),
-        # Model(
-        #     id="minimax-02-turbo",
-        #     name="MiniMax Speech-02-Turbo",
-        #     model_type=ModelType.TTS,
-        #     is_open=False,
-        #     model_url="http://minimax.io/",
-        # ),
-        # Model(
-        #     id="lanternfish-1",
-        #     name="Lanternfish",
-        #     model_type=ModelType.TTS,
-        #     is_open=False,
-        # ),
+        )
     ]
     conversational_models = [
         Model(

tts.py

@@ -47,7 +47,7 @@ data = {"text": "string", "provider": "string", "model": "string"}
 
 def predict_index_tts(text, reference_audio_path=None):
     from gradio_client import Client, handle_file
-    client = Client("kemuriririn/IndexTTS", hf_token=os.getenv("HF_TOKEN"))
+    client = Client("kemuriririn/IndexTTS", hf_token=get_zerogpu_token())
     if reference_audio_path:
         prompt = handle_file(reference_audio_path)
     else:
@@ -65,7 +65,7 @@ def predict_index_tts(text, reference_audio_path=None):
 
 def predict_spark_tts(text, reference_audio_path=None):
     from gradio_client import Client, handle_file
-    client = Client("kemuriririn/SparkTTS", hf_token=os.getenv("HF_TOKEN"))
+    client = Client("kemuriririn/SparkTTS", hf_token=get_zerogpu_token())
     prompt_wav = None
     if reference_audio_path:
         prompt_wav = handle_file(reference_audio_path)