Update app.py

app.py

@@ -2,7 +2,7 @@ import os
 import subprocess
 import tempfile
 
-subprocess.run('pip install flash-attn==2.8.0.post2 --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
+#subprocess.run('pip install flash-attn==2.8.0.post2 --no-build-isolation', env={'FLASH_ATTENTION_SKIP_CUDA_BUILD': "TRUE"}, shell=True)
 
 import threading
 
@@ -56,7 +56,7 @@ def load_model():
         MODEL_ID,
         dtype=torch.bfloat16,  # if device == "cuda" else torch.float32,
         device_map="auto",  # if device == "cuda" else None,
-        attn_implementation="flash_attention_2",# "kernels-community/vllm-flash-attn3", #  #
+        attn_implementation="kernels-community/flash-attn" # "flash_attention_2",# "kernels-community/vllm-flash-attn3", #  #
     )  # .cuda()
     print(f"Selected device:", device)
     return model, tokenizer, processor, device