update multiprocessing handling + concurency count

app_dialogue.py

@@ -634,24 +634,19 @@ def model_inference(
     query = prompt_list_to_tgi_input(formated_prompt_list)
     all_meme_images = []
     with concurrent.futures.ThreadPoolExecutor(max_workers=4) as executor:
-        futures = [
-            executor.submit(
+        all_meme_images = list(
+            executor.map(
                 generate_meme,
-                client,
-                query,
-                image,
-                font_meme_text,
-                all_caps_meme_text,
-                text_at_the_top,
-                generation_args,
+                [client for _ in range(4)],
+                [query for _ in range(4)],
+                [image for _ in range(4)],
+                [font_meme_text for _ in range(4)],
+                [all_caps_meme_text for _ in range(4)],
+                [text_at_the_top for _ in range(4)],
+                [generation_args for _ in range(4)],
             )
-            for i in range(4)
-        ]
-
-        for future in concurrent.futures.as_completed(futures):
-            meme_image = future.result(timeout=45)
-            if meme_image:
-                all_meme_images.append(meme_image)
+        )
+    all_meme_images = [meme for meme in all_meme_images if meme is not None]
     return user_prompt_str, all_meme_images, chat_history
 
 
@@ -941,5 +936,5 @@ with gr.Blocks(title="AI Meme Generator", theme=gr.themes.Base(), css=css) as de
         # outputs=[template_gallery],
         queue=False,
     )
-demo.queue(concurrency_count=4, max_size=40, api_open=False)
+demo.queue(concurrency_count=8, max_size=40, api_open=False)
 demo.launch(max_threads=400)