Update app.py

app.py

@@ -14,11 +14,32 @@ def model_choices():
 
 model_description = {model['filename']: model['description'] for model in data_json}
 
+
 def llm_intro(selected_model):
     return model_description.get(selected_model, "No description available for this model selection.")
 
+model_cache = {}  # Global cache
+
+
+def load_model_with_cache(model_name):
+    """
+    This function checks the cache before loading a model.
+    If the model is cached, it returns the cached version.
+    Otherwise, it loads the model, caches it, and then returns it.
+    """
+    if model_name not in model_cache:
+        # Load the model. This is a placeholder; replace with actual model loading logic.
+        model = GPT4All(model_name)  # Placeholder function
+        model_cache[model_name] = model
+    return model_cache[model_name]
+
+
 def generate_text(input_text, selected_model):
-    model = GPT4All(selected_model)
+    """
+    Generate text using the selected model.
+    This function now uses the caching mechanism to load models.
+    """
+    model = load_model_with_cache(selected_model)
     output = model.generate(input_text, max_tokens=100)
     return output