Update app.py

app.py

@@ -67,11 +67,12 @@ def Generate(image_input, prompt, negative_prompt, strength, guidance_scale, num
 	return image, f"{minutes:02d}:{seconds:02d}"
 def Loading(model):
 	global text2img, img2img
-	text2img = StableDiffusionPipeline.from_pretrained(model, torch_dtype=torch.float16).to(device)
+	text2img = StableDiffusionPipeline.from_pretrained(model, torch_dtype=torch.float16, use_safetensors=True).to(device)
 	text2img.safety_checker = None
 	text2img.scheduler = EulerDiscreteScheduler.from_config(text2img.scheduler.config)
-	text2img.enable_xformers_memory_efficient_attention()
-	text2img.vae.enable_xformers_memory_efficient_attention()
+	if device == "cuda":
+		text2img.enable_xformers_memory_efficient_attention()
+		text2img.vae.enable_xformers_memory_efficient_attention()
 	img2img = StableDiffusionImg2ImgPipeline(**text2img.components)
 	return model
 with gr.Blocks() as demo:
@@ -104,4 +105,4 @@ with gr.Blocks() as demo:
 	generate.click(Generate, [image_input, prompt, negative_prompt, strength, guidance_scale, num_inference_steps, width, height, seed], [image_output, text_output])
 	loading.click(Loading, model, model)
 	set_language.change(update_language, set_language, [model, loading, image_input, prompt, negative_prompt, generate, strength, guidance_scale, num_inference_steps, width, height, seed])
-demo.queue().launch()
+demo.queue(concurrency_count=24, max_size=32).launch(max_threads=128)