Update app.py

app.py

@@ -2,20 +2,23 @@ import gradio as gr
 import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
-# Replace with your actual merged model repo
-model_id = "lakshraina2/leetcode-coder-1.5B" 
+# The public model ID
+model_id = "lakshraina2/leetcodeAI" 
 
-print("Loading model on CPU...")
-tokenizer = AutoTokenizer.from_pretrained(model_id)
+print("Loading model on CPU (Public Access)...")
+
+# Force token=False to bypass the 401 error on public repos
+tokenizer = AutoTokenizer.from_pretrained(model_id, token=False)
 model = AutoModelForCausalLM.from_pretrained(
     model_id, 
-    torch_dtype=torch.float32, # CPU needs float32
-    device_map={"": "cpu"}     # Force CPU
+    torch_dtype=torch.float32, 
+    device_map={"": "cpu"},
+    token=False # This is the magic fix
 )
 
 def solve(problem_text):
     prompt = f"### Instruction:\nSolve this LeetCode problem:\n{problem_text}\n\n### Response:\n"
-    inputs = tokenizer(prompt, return_tensors="pt") # No .to("cuda")!
+    inputs = tokenizer(prompt, return_tensors="pt")
     
     with torch.no_grad():
         outputs = model.generate(
@@ -28,6 +31,5 @@ def solve(problem_text):
     solution = tokenizer.decode(outputs[0], skip_special_tokens=True)
     return solution.split("### Response:\n")[-1].strip()
 
-# Gradio 4 interface
 iface = gr.Interface(fn=solve, inputs="text", outputs="text")
 iface.launch()