Upload app.py with huggingface_hub

app.py

@@ -1,11 +1,11 @@
 import os
+os.environ.setdefault("CHATTERBOX_FLASH_ENGINE", "torch")
 os.environ.setdefault("NUMBA_DISABLE_CUDA", "1")
 
 import spaces
 import torch
 print(f"After import spaces: cuda_available={torch.cuda.is_available()}")
 
-# Install chatterbox packages
 import subprocess, sys
 subprocess.run(
     [sys.executable, "-m", "pip", "install", "--no-deps",
@@ -17,11 +17,17 @@ print(f"After pip install: cuda_available={torch.cuda.is_available()}")
 from chatterbox_flash import ChatterboxFlashTTS
 print(f"After chatterbox import: cuda_available={torch.cuda.is_available()}")
 
+MODEL_ID = "ResembleAI/chatterbox-flash"
+print(f"Loading model from {MODEL_ID}...")
+tts = ChatterboxFlashTTS.from_pretrained(MODEL_ID, device="cuda", dtype=torch.bfloat16)
+print(f"After model load: cuda_available={torch.cuda.is_available()}")
+print("Model loaded successfully.")
+
 import gradio as gr
 
 @spaces.GPU(duration=60)
 def test_gpu(text):
-    return f"CUDA available: {torch.cuda.is_available()}, device count: {torch.cuda.device_count()}, device name: {torch.cuda.get_device_name(0)}"
+    return f"CUDA: {torch.cuda.is_available()}, devices: {torch.cuda.device_count()}"
 
 demo = gr.Interface(fn=test_gpu, inputs=gr.Textbox(value="test"), outputs=gr.Textbox())
 demo.launch(mcp_server=True)