Updated app.py

app.py

@@ -1,8 +1,12 @@
 import streamlit as st
+import transformers
 from transformers import AutoTokenizer, AutoModelForCausalLM
 import torch
-
-
+# device = "cpu"
+# if torch.cuda.is_available():
+#     device = "cuda"
+# if torch.backends.mps.is_available():
+#     device = "mps"
 
 # torch.set_default_device("cuda")
 # Load the Phi 2 model and tokenizer
@@ -13,10 +17,11 @@ tokenizer = AutoTokenizer.from_pretrained(
 
 model = AutoModelForCausalLM.from_pretrained(
     "microsoft/phi-2",
-    device_map="auto",
+    device_map="cpu",
     trust_remote_code=True,
     # offload_folder="offload",
     torch_dtype=torch.float32
+    # torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
 )
 
 # Streamlit UI
@@ -28,13 +33,15 @@ prompt = st.text_area("Enter your prompt:", """Write a short summary about how t
 # Generate output based on user input
 if st.button("Generate Output"):
     with torch.no_grad():
-        token_ids = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt",  return_attention_mask=False)
+        token_ids = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt",  
+                                     # return_attention_mask=False
+                                    )
         output_ids = model.generate(
             token_ids.to(model.device),
             max_new_tokens=512,
             do_sample=True,
             temperature=0.3,
-            max_length=200
+            # max_length=200
         )
 
     output = tokenizer.decode(output_ids[0][token_ids.size(1):])