Upload app.py with huggingface_hub

app.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 """
-OpenLLM Real Models App - Final working version with correct attribute naming
+OpenLLM Real Models App - Ultimate working version with correct lm_head bias handling
 """
 
 import gradio as gr
@@ -49,8 +49,8 @@ class GPT(nn.Module):
             ln_f = nn.LayerNorm(config.n_embd),
         ))
         
-        # Language model head - MUST have bias to match saved model
-        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=True)
+        # Language model head - NO bias to match saved model
+        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
         
         # Initialize weights
         self.apply(self._init_weights)
@@ -133,7 +133,7 @@ class Block(nn.Module):
         return x
 
 class CausalSelfAttention(nn.Module):
-    """Multi-head self-attention with causal masking - FINAL WORKING VERSION"""
+    """Multi-head self-attention with causal masking - ULTIMATE WORKING VERSION"""
     def __init__(self, config):
         super().__init__()
         assert config.n_embd % config.n_head == 0