Update app.py

app.py

@@ -8,12 +8,13 @@ MODEL_ID = "Cydercoder/qwen2.5-coder-3b"
 
 print("Loading tokenizer...")
 tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
-print("Loading model on CPU (Free Tier)...")
-# We load in 8-bit or float32 to fit inside the free 16GB CPU RAM limit safely
+
+print("Loading model on CPU via standard clean mapping...")
+# Forcing float32 without low_cpu_mem_usage prevents the Transformers v5 thread crash
 model = AutoModelForCausalLM.from_pretrained(
     MODEL_ID,
     torch_dtype=torch.float32,
-    low_cpu_mem_usage=True
+    device_map="cpu"
 )
 
 def chat_function(message, history):
@@ -51,7 +52,7 @@ def chat_function(message, history):
         partial_text += new_text
         yield partial_text
 
-# 4. Initialize the gorgeous, interactive native Gradio browser interface
+# 4. Initialize the native Gradio browser interface
 demo = gr.ChatInterface(
     fn=chat_function,
     title="🤖 Cydercoder Qwen 3B AI Chatbot",