Raise max new tokens slider to 7072

app.py

@@ -20,6 +20,7 @@ MODELS = {
 }
 
 HF_TOKEN = os.environ.get("HF_TOKEN")
+MAX_NEW_TOKENS = 8096 - 1024  # leave room for the prompt in ~8k context
 
 tokenizers: dict[str, AutoTokenizer] = {}
 models: dict[str, AutoModelForCausalLM] = {}
@@ -138,7 +139,7 @@ with gr.Blocks(title="Tiny Aya Compare") as demo:
 
     with gr.Row():
         max_new_tokens = gr.Slider(
-            16, 512, value=256, step=16, label="Max new tokens"
+            16, MAX_NEW_TOKENS, value=256, step=16, label="Max new tokens"
         )
         temperature = gr.Slider(0, 1.5, value=0.7, step=0.05, label="Temperature")
         top_p = gr.Slider(0.1, 1.0, value=0.9, step=0.05, label="Top-p")