Update app.py

app.py

@@ -3,20 +3,23 @@ from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
 import torch
 import requests
 
-# Load DeepSeek-R1 model
+# Load DeepSeek-R1 model with trust_remote_code enabled
 model_name = "deepseek-ai/DeepSeek-R1"
 tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
 
-# Ensure the model uses float16 instead of fp8
-model = AutoModelForCausalLM.from_pretrained(
-    model_name,
-    torch_dtype=torch.float16,  # Forces float16 to prevent fp8 issue
-    device_map="auto",
-    trust_remote_code=True
-)
+# Ensure compatibility with `flash_attn` and force proper dtype
+try:
+    model = AutoModelForCausalLM.from_pretrained(
+        model_name,
+        torch_dtype=torch.float16,  # Forces float16 to prevent fp8 issue
+        device_map="auto",
+        trust_remote_code=True
+    )
+except ImportError as e:
+    raise RuntimeError("Missing required dependency: flash_attn. Install with `pip install flash_attn`") from e
 
 # Use a text-generation pipeline for better inference
-generator = pipeline("text-generation", model=model, tokenizer=tokenizer, device=0)
+generator = pipeline("text-generation", model=model, tokenizer=tokenizer, device=0 if torch.cuda.is_available() else -1)
 
 
 # 🎯 Function to extract interests from user input