.

app.py

@@ -17,7 +17,7 @@ snapshot_download(repo_id=repo_name, local_dir=".", allow_patterns=model_name)
 
 model = Llama(
     model_path=model_name,
-    n_ctx=2000,
+    n_ctx=2048,
     n_parts=1,
 )
 
@@ -29,7 +29,7 @@ def generate(
     temperature=0.1,
     top_p=0.75,
     top_k=40,
-    max_new_tokens=128,
+    max_new_tokens=512,
 ):
     prompt = template.format(input)
     output = model(prompt, 
@@ -52,7 +52,7 @@ g = gr.Interface(
         gr.components.Slider(minimum=0, maximum=1, value=0.8, label="Top p"),
         gr.components.Slider(minimum=0, maximum=100, step=1, value=50, label="Top k"),
         gr.components.Slider(
-            minimum=1, maximum=1024, step=1, value=256, label="Max tokens"
+            minimum=1, maximum=2048, step=1, value=512, label="Max tokens"
         ),
     ],
     outputs=[
@@ -64,5 +64,5 @@ g = gr.Interface(
     title="tinyllama-1.1b-chat gguf",
     description=''
 )
-g.queue(concurrency_count=1)
+g.queue(concurrency_count=2)
 g.launch()