fix: add missing custom_model_metadata state variable and update get_model_info to handle custom models

app.py

@@ -901,16 +901,56 @@ def calculate_effective_max_tokens(model_key: str, max_tokens: int, enable_reaso
     return max_tokens
 
 
-def get_model_info(model_key: str, n_threads: int = 2) -> Tuple[str, str, float, int]:
+def get_model_info(model_key: str, n_threads: int = 2, custom_metadata: Optional[dict] = None) -> Tuple[str, str, float, int]:
     """Get model information and inference settings for UI display.
 
     Args:
         model_key: Model identifier from AVAILABLE_MODELS
         n_threads: Number of CPU threads currently configured
+        custom_metadata: Optional metadata for custom models (repo_id, filename, size_mb)
 
     Returns:
         Tuple of (info_text, temperature, top_p, top_k)
     """
+    # Handle custom model case
+    if model_key == "custom_hf" and custom_metadata:
+        repo_id = custom_metadata.get("repo_id", "Unknown")
+        filename = custom_metadata.get("filename", "Unknown")
+        size_mb = custom_metadata.get("size_mb", 0)
+        size_str = f"{size_mb:.1f} MB" if size_mb > 0 else "Unknown"
+        
+        # Determine thread preset label
+        if n_threads == 2:
+            thread_label = "HF Free Tier"
+        elif n_threads == 8:
+            thread_label = "HF Upgrade Tier"
+        else:
+            thread_label = "Custom"
+        
+        info_text = (
+            f"## 🤖 Custom GGUF Model\n\n"
+            f"### 📊 Model Specs\n"
+            f"| Property | Value |\n"
+            f"|----------|-------|\n"
+            f"| **Repository** | `{repo_id}` |\n"
+            f"| **Quantization** | `{filename}` |\n"
+            f"| **Size** | {size_str} |\n"
+            f"| **Context** | Dynamic (up to 32K) |\n\n"
+            f"### 🖥️ Hardware Configuration\n"
+            f"| Property | Value |\n"
+            f"|----------|-------|\n"
+            f"| **CPU Threads** | {n_threads} ({thread_label}) |\n\n"
+            f"### ⚙️ Inference Settings\n"
+            f"| Property | Value |\n"
+            f"|----------|-------|\n"
+            f"| **Temperature** | 0.6 |\n"
+            f"| **Top P** | 0.9 |\n"
+            f"| **Top K** | 40 |\n"
+            f"| **Repeat Penalty** | 1.0 |"
+        )
+        return info_text, "0.6", 0.9, 40
+    
+    # Handle predefined models
     m = AVAILABLE_MODELS[model_key]
     usable_ctx = min(m["max_context"], MAX_USABLE_CTX)
     settings = m["inference_settings"]
@@ -1654,6 +1694,13 @@ def create_interface():
                     
                     # Hidden state to store loaded custom model
                     custom_model_state = gr.State(value=None)
+                    
+                    # Hidden state to store custom model metadata (repo_id, filename, size)
+                    custom_model_metadata = gr.State(value={
+                        "repo_id": None,
+                        "filename": None,
+                        "size_mb": 0,
+                    })
                 
                 # Model info section (dynamic)
                 with gr.Group():
@@ -1708,7 +1755,7 @@ def create_interface():
                         )
         
         # Function to update settings when model changes
-        def update_settings_on_model_change(model_key, thread_config, custom_threads):
+        def update_settings_on_model_change(model_key, thread_config, custom_threads, custom_metadata=None):
             """Update all Advanced Settings when model selection changes."""
             # Calculate n_threads based on preset
             thread_preset_map = {
@@ -1718,7 +1765,7 @@ def create_interface():
             }
             n_threads = thread_preset_map.get(thread_config, 2)
 
-            info_text, temp_str, top_p_val, top_k_val = get_model_info(model_key, n_threads=n_threads)
+            info_text, temp_str, top_p_val, top_k_val = get_model_info(model_key, n_threads=n_threads, custom_metadata=custom_metadata)
             temperature = float(temp_str) if temp_str else 0.6
             return temperature, top_p_val, top_k_val, info_text