Update app.py

app.py

@@ -5,15 +5,16 @@ import torch
 # Set the random seed for reproducibility
 torch.random.manual_seed(0)
 
-# Load the model and tokenizer
-model_name = "microsoft/Phi-3.5-mini-instruct"
+# Load the model and tokenizer with pinned revision
 model = AutoModelForCausalLM.from_pretrained(
-    model_name, 
+    "microsoft/Phi-3.5-mini-instruct", 
     device_map="auto", 
     torch_dtype="auto", 
-    trust_remote_code=True
+    trust_remote_code=True,
+    revision="main",  # Pin the revision for stability
+    attn_implementation="eager"  # Use eager attention implementation if flash-attention is not installed
 )
-tokenizer = AutoTokenizer.from_pretrained(model_name)
+tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3.5-mini-instruct")
 
 # Set up the pipeline
 pipe = pipeline(