Update app.py

app.py

@@ -14,9 +14,7 @@ print("Installing flash-attn...")
 subprocess.run(
     "pip install flash-attn --no-build-isolation",
     env={"FLASH_ATTENTION_SKIP_CUDA_BUILD": "TRUE"},
-    shell=True,
-    capture_output=True,  # Capture output for debugging
-    text=True  # Decode output as text
+    shell=True
 )
 
 from huggingface_hub import snapshot_download
@@ -76,11 +74,9 @@ device = "cuda:0"
 
 model = AutoModelForCausalLM.from_pretrained(
     "m-a-p/YuE-s1-7B-anneal-en-cot",
-    torch_dtype=torch.bfloat16,
+    torch_dtype=torch.float16,
     attn_implementation="flash_attention_2",
-    # quantization_config=quantization_config,
     low_cpu_mem_usage=True,
-    # device_map="auto"     
 ).to(device)
 model.eval()