Update app.py

app.py

@@ -176,6 +176,18 @@ def caption_video(video_path: str) -> str:
 
     T = Timer()
     model, processor = _load_model_and_processor()
+
+    print("[ENV] MAX_NEW_TOKENS =", MAX_NEW_TOKENS, flush=True)
+    print("[CUDA] available =", torch.cuda.is_available(), flush=True)
+    if torch.cuda.is_available():
+        print("[CUDA] device =", torch.cuda.get_device_name(0), flush=True)
+    try:
+        print("[MODEL] device_map =", getattr(model, "hf_device_map", None), flush=True)
+        print("[MODEL] first_param_device =", next(model.parameters()).device, flush=True)
+    except Exception as e:
+        print("[MODEL] device inspect error:", e, flush=True)    
+    
+    
     # 1) 可能轉碼 / 降維 / 限 FPS
     t = time.perf_counter()
     safe_path, used_temp, tr_info = maybe_transcode(video_path)
@@ -220,6 +232,9 @@ def caption_video(video_path: str) -> str:
         temperature=0.0,
         top_p=1.0
     )
+    
+    print("[ENV] MAX_NEW_TOKENS =", MAX_NEW_TOKENS, flush=True)
+    
     t = time.perf_counter()
     with torch.inference_mode():
         generated_ids = model.generate(**inputs, **gen_kwargs)