Update app.py

app.py

@@ -17,31 +17,6 @@ from transformers import (
 
 print("\n=== Environment Setup ===")
 
-if torch.cuda.is_available():
-    print(f"GPU detected: {torch.cuda.get_device_name(0)}")
-    try:
-        subprocess.run(
-            "pip install flash-attn --no-build-isolation",
-            shell=True,
-            check=True,
-        )
-        print("✅ flash-attn installed successfully")
-    except subprocess.CalledProcessError as e:
-        print("⚠️ flash-attn installation failed:", e)
-else:
-    print("⚙️ CPU detected — skipping flash-attn installation")
-    # Disable flash-attn references safely
-    os.environ["DISABLE_FLASH_ATTN"] = "1"
-    os.environ["FLASH_ATTENTION_SKIP_CUDA_BUILD"] = "TRUE"
-try:
-    from transformers.utils import import_utils
-
-    if "flash_attn" not in import_utils.PACKAGE_DISTRIBUTION_MAPPING:  # type: ignore
-        print(import_utils.PACKAGE_DISTRIBUTION_MAPPING)
-        import_utils.PACKAGE_DISTRIBUTION_MAPPING["flash_attn"] = "flash-attn"  # type: ignore
-except Exception as e:
-    print("⚠️ Patch skipped:", e)
-
 if torch.cuda.is_available():
     device = torch.device("cuda")
     print(f"Using GPU: {torch.cuda.get_device_name(device)}")
@@ -50,11 +25,6 @@ else:
     print("Using CPU")
 
 print("\n=== Model Loading ===")
-import torch
-import transformers
-
-from transformers.utils.import_utils import is_flash_attn_2_available
-print("is_flash_attn_2_available: ", is_flash_attn_2_available())
 
 chat_model_name = "sapienzanlp/Minerva-7B-instruct-v1.0"
 cls_model_name = "saiteki-kai/QA-DeBERTa-v3-large-binary-3"