Update app.py

app.py

@@ -2,13 +2,32 @@ import gradio as gr
 import random
 from llama_cpp import Llama
 
-# Charger le modèle
-llm = Llama(
-    model_path="model/WizzGPTv7-q8_0.gguf",
-    n_threads=2,
-    n_ctx=1024,
-    use_mlock=True
-)
+# Liste des modèles disponibles
+model_versions = {
+    "WizzGPTv2": "model/WizzGPTv2-q8_0.gguf",
+    "WizzGPTv3": "model/WizzGPTv3-q8_0.gguf",
+    "WizzGPTv4": "model/WizzGPTv4-q8_0.gguf",
+    "WizzGPTv5": "model/WizzGPTv5-q8_0.gguf",
+    "WizzGPTv6": "model/WizzGPTv6-q8_0.gguf",
+    "WizzGPTv7": "model/WizzGPTv7-q8_0.gguf"  # Par défaut
+}
+
+# Modèle initial (v7 par défaut)
+llm = None
+current_model = "WizzGPTv7"
+
+def load_model(model_name):
+    global llm, current_model
+    current_model = model_name
+    llm = Llama(
+        model_path=model_versions[model_name],
+        n_threads=2,
+        n_ctx=1024,
+        use_mlock=True
+    )
+
+# Charger modèle initial
+load_model(current_model)
 
 # Liste complète des débuts de prompt
 prompt_list = [
@@ -285,7 +304,7 @@ prompt_list = [
     "A visually arresting photographic study of "
 ]
 
-# Generation function
+# Fonction de génération
 def generate_completions(prompt, n_responses, max_tokens, temperature, top_p, top_k, repeat_penalty):
     results = []
     for _ in range(n_responses):
@@ -302,7 +321,7 @@ def generate_completions(prompt, n_responses, max_tokens, temperature, top_p, to
         results.append(f"{prompt}{text}")
     return "\n\n".join(results)
 
-# UI logic
+# Fonctions UI
 def set_random_prompt(n):
     selected = random.choice(prompt_list)
     return selected, f"🎲 Random Prompt{'s' if n > 1 else ''}", f"🚀 Generate Prompt{'s' if n > 1 else ''}"
@@ -310,22 +329,21 @@ def set_random_prompt(n):
 def set_prompt_from_dropdown(choice, n):
     return choice, f"🎲 Random Prompt{'s' if n > 1 else ''}", f"🚀 Generate Prompt{'s' if n > 1 else ''}"
 
-# Gradio Interface
-with gr.Blocks(css=".gr-button { font-size: 16px; }") as demo:
-    gr.Markdown("""
-    # ✨ WizzGPTv7 — Text Completion (GGUF - CPU)
+def on_model_change(model_name):
+    load_model(model_name)
+    return f"Model loaded: {model_name}"
 
-    **Lightweight GGUF model (Q8_0), optimized for CPU.**  
-    Works in *completion* mode (not chatbot), returning full prompts + outputs.
+# Interface Gradio
+with gr.Blocks(css=".gr-button { font-size: 16px; }") as demo:
+    gr.Markdown("## ✨ WizzGPT — Text Completion (GGUF - CPU)")
 
-    - Click `🎲 Random Prompt` for a random start
-    - Use the dropdown to pick from hundreds of prompt starters
-    - Adjust generation settings with the sliders below
-    """)
+    with gr.Row():
+        model_selector = gr.Dropdown(choices=list(model_versions.keys()), value="WizzGPTv7", label="🧠 Select Model")
+        model_status = gr.Textbox(value=f"Model loaded: {current_model}", label="Status", interactive=False)
 
     with gr.Group():
         with gr.Row():
-            prompt_input = gr.Textbox(label="📝 Prompt", lines=3, placeholder="Type a prompt or pick one from the list...")
+            prompt_input = gr.Textbox(label="📝 Prompt", lines=3, placeholder="Type or pick a prompt...")
         with gr.Row():
             prompt_dropdown = gr.Dropdown(choices=prompt_list, label="📌 Prompt Examples", scale=2)
             random_button = gr.Button("🎲 Random Prompt", scale=1)
@@ -335,26 +353,20 @@ with gr.Blocks(css=".gr-button { font-size: 16px; }") as demo:
 
     with gr.Accordion("⚙️ Advanced Settings", open=False):
         max_tokens = gr.Slider(10, 200, value=75, step=5, label="Max Tokens")
-        temperature = gr.Slider(0.1, 2.0, value=1.1, step=0.05, label="Temperature")
+        temperature = gr.Slider(0.1, 2.0, value=1.25, step=0.05, label="Temperature")
         top_p = gr.Slider(0.0, 1.0, value=0.90, step=0.05, label="Top-p")
-        top_k = gr.Slider(0, 128, value=40, step=1, label="Top-k")
-        repeat_penalty = gr.Slider(0.5, 2.0, value=1.32, step=0.05, label="Repeat Penalty")
+        top_k = gr.Slider(0, 128, value=128, step=1, label="Top-k")
+        repeat_penalty = gr.Slider(0.5, 2.0, value=1.4, step=0.05, label="Repeat Penalty")
 
     generate_button = gr.Button("🚀 Generate Prompt(s)")
     output = gr.Textbox(label="📄 Completions", lines=12)
 
-    # Reactive logic
+    # Interactions
     prompt_dropdown.change(set_prompt_from_dropdown, [prompt_dropdown, n_responses], [prompt_input, random_button, generate_button])
     random_button.click(set_random_prompt, inputs=n_responses, outputs=[prompt_input, random_button, generate_button])
-    n_responses.change(
-        lambda n: (f"🎲 Random Prompt{'s' if n > 1 else ''}", f"🚀 Generate Prompt{'s' if n > 1 else ''}"),
-        inputs=n_responses,
-        outputs=[random_button, generate_button]
-    )
-    generate_button.click(
-        fn=generate_completions,
-        inputs=[prompt_input, n_responses, max_tokens, temperature, top_p, top_k, repeat_penalty],
-        outputs=output
-    )
+    n_responses.change(lambda n: (f"🎲 Random Prompt{'s' if n > 1 else ''}", f"🚀 Generate Prompt{'s' if n > 1 else ''}"), inputs=n_responses, outputs=[random_button, generate_button])
+    generate_button.click(fn=generate_completions, inputs=[prompt_input, n_responses, max_tokens, temperature, top_p, top_k, repeat_penalty], outputs=output)
+
+    model_selector.change(on_model_change, inputs=model_selector, outputs=model_status)
 
 demo.launch()