cpu_gpu_mps

app.py

@@ -1,12 +1,28 @@
 import gradio as gr
-
 import torch
+
 from diffusers import DiffusionPipeline
 from diffusers import EulerDiscreteScheduler
 
+device = "cpu"
+dtype = torch.float32
+if torch.cuda.is_available():
+    device = "cuda"
+    dtype = torch.float16
+
+# check if MPS is available OSX only M1/M2/M3 chips
+mps_available = hasattr(torch.backends, "mps") and torch.backends.mps.is_available()
+if mps_available:
+    device = "mps"
+    dtype = torch.float16
+#print(f"device: {device}, dtype: {dtype}")
 
-pipeline = DiffusionPipeline.from_pretrained("recoilme/ColorfulXL-Lightning",variant="fp16"#, torch_dtype=torch.float16
-                                             , use_safetensors=True)#.to("cuda")
+
+pipeline = DiffusionPipeline.from_pretrained("recoilme/ColorfulXL-Lightning",
+                                             variant="fp16",
+                                             torch_dtype=dtype,
+                                             use_safetensors=True)
+pipeline.to(device)
 pipeline.scheduler = EulerDiscreteScheduler.from_config(pipeline.scheduler.config, timestep_spacing="trailing")
 
 
@@ -17,28 +33,22 @@ with gr.Blocks() as interface:
         with gr.Column():
             with gr.Row():
                 with gr.Column():
-                    prompt = gr.Textbox(label="Prompt", info="What do you want?", value="girl sitting on a small hill looking at night sky, back view, distant exploding moon, nights darkness, intricate circuits and sensors, photographic realism style, detailed textures, peacefulness, mysterious.", lines=4, interactive=True)
+                    prompt = gr.Textbox(label="Prompt", info="What do you want?", value="girl sitting on a small hill looking at night sky, back view, distant exploding moon", lines=4, interactive=True)
                 with gr.Column():
                     generate_button = gr.Button("Generate")
-                    output = gr.Image()
-            with gr.Row():
-                with gr.Accordion(label="Advanced Settings", open=False):
-                    with gr.Row():
-                        with gr.Column():
-                            width = gr.Slider(label="Width", info="The width in pixels of the generated image.", value=576, minimum=512, maximum=1280, step=64, interactive=True)
-                            height = gr.Slider(label="Height", info="The height in pixels of the generated image.", value=832, minimum=512, maximum=1280, step=64, interactive=True)
-                        with gr.Column():
-                            sampling_steps = gr.Slider(label="Sampling Steps", info="The number of denoising steps.", value=5, minimum=3, maximum=10, step=1, interactive=True)
-    
+                    with gr.Accordion(label="Advanced Settings", open=False):
+                        with gr.Row():
+                            with gr.Column():
+                                width = gr.Slider(label="Width", info="The width in pixels of the generated image.", value=576, minimum=512, maximum=1280, step=64, interactive=True)
+                                height = gr.Slider(label="Height", info="The height in pixels of the generated image.", value=832, minimum=512, maximum=1280, step=64, interactive=True)
+                            with gr.Column():
+                                sampling_steps = gr.Slider(label="Sampling Steps", info="The number of denoising steps.", value=5, minimum=3, maximum=10, step=1, interactive=True)
+                        
             with gr.Row():
-                about_text = """
-                Based on: Stable Diffusion XL Image Generation interface built by Noa Roggendorff.
-                
-                You can enter a prompt and negative prompt, adjust the image size and sampling steps, and click the "Generate" button to generate an image.
-                """
-                gr.Markdown(about_text)
+                output = gr.Image()
         
         generate_button.click(fn=generate, inputs=[prompt, width, height, sampling_steps], outputs=[output])
 
 if __name__ == "__main__":
     interface.launch()
+