Update app.py

app.py

@@ -1,33 +1,37 @@
 import gradio as gr
 import numpy as np
 import random
+import os  # Import os to access secrets
 from diffusers import DiffusionPipeline
 import torch
 
-# Device setup - CPU for now
-device = "cpu"
+# Set device
+device = "cuda" if torch.cuda.is_available() else "cpu"
 
-# Model setup (replace with the model you want to use)
+# Use Hugging Face secret from your space environment
+access_token = os.getenv('HUGGINGFACE_TOKEN')
+
+# Replace with the correct model repo ID
 model_repo_id = "KingNish/Realtime-FLUX"
 
-# CPU runs with float32 as float16 is typically for GPU
-torch_dtype = torch.float32
+# Set the torch dtype based on availability of CUDA
+if torch.cuda.is_available():
+    torch_dtype = torch.float16
+else:
+    torch_dtype = torch.float32
 
-# Load the model on CPU
-pipe = DiffusionPipeline.from_pretrained(model_repo_id, torch_dtype=torch_dtype).to(device)
+# Load the pipeline with the access token
+pipe = DiffusionPipeline.from_pretrained(model_repo_id, torch_dtype=torch_dtype, use_auth_token=access_token).to(device)
 
-# Constants
 MAX_SEED = np.iinfo(np.int32).max
-MAX_IMAGE_SIZE = 512  # Reduce size for CPU performance
+MAX_IMAGE_SIZE = 1024
 
 def infer(prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps, progress=gr.Progress(track_tqdm=True)):
-
     if randomize_seed:
         seed = random.randint(0, MAX_SEED)
-        
-    generator = torch.manual_seed(seed)
     
-    # Generate image using the pipeline
+    generator = torch.Generator().manual_seed(seed)
+    
     image = pipe(
         prompt=prompt, 
         negative_prompt=negative_prompt,
@@ -40,14 +44,12 @@ def infer(prompt, negative_prompt, seed, randomize_seed, width, height, guidance
     
     return image, seed
 
-# Sample prompts
 examples = [
-    "Astronaut in a jungle, cold color palette, detailed, 8k",
-    "A robot playing a guitar in a futuristic city",
-    "A glowing blue cat sitting on a floating island"
+    "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k",
+    "An astronaut riding a green horse",
+    "A delicious ceviche cheesecake slice",
 ]
 
-# Gradio Interface
 css = """
 #col-container {
     margin: 0 auto;
@@ -56,84 +58,26 @@ css = """
 """
 
 with gr.Blocks(css=css) as demo:
-    
     with gr.Column(elem_id="col-container"):
-        gr.Markdown("# Realtime-FLUX Text-to-Image Generator")
-
+        gr.Markdown("# Text-to-Image Gradio Template")
         with gr.Row():
-            prompt = gr.Textbox(
-                label="Prompt",
-                show_label=False,
-                max_lines=1,
-                placeholder="Describe the image you want to generate...",
-                container=False,
-            )
-            
-            run_button = gr.Button("Generate")
-
-        result = gr.Image(label="Generated Image", show_label=False)
+            prompt = gr.Text(label="Prompt", show_label=False, max_lines=1, placeholder="Enter your prompt", container=False)
+            run_button = gr.Button("Run", scale=0)
+        result = gr.Image(label="Result", show_label=False)
 
         with gr.Accordion("Advanced Settings", open=False):
-            negative_prompt = gr.Textbox(
-                label="Negative Prompt",
-                max_lines=1,
-                placeholder="Enter a negative prompt (optional)",
-                visible=True,
-            )
-            
-            seed = gr.Slider(
-                label="Seed",
-                minimum=0,
-                maximum=MAX_SEED,
-                step=1,
-                value=0,
-            )
-            
-            randomize_seed = gr.Checkbox(label="Randomize Seed", value=True)
-
+            negative_prompt = gr.Text(label="Negative prompt", max_lines=1, placeholder="Enter a negative prompt", visible=False)
+            seed = gr.Slider(label="Seed", minimum=0, maximum=MAX_SEED, step=1, value=0)
+            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
             with gr.Row():
-                width = gr.Slider(
-                    label="Width",
-                    minimum=256,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=384,  # Optimal for CPU
-                )
-                height = gr.Slider(
-                    label="Height",
-                    minimum=256,
-                    maximum=MAX_IMAGE_SIZE,
-                    step=32,
-                    value=384,  # Optimal for CPU
-                )
-            
-            with gr.Row():
-                guidance_scale = gr.Slider(
-                    label="Guidance Scale",
-                    minimum=0.0,
-                    maximum=10.0,
-                    step=0.5,
-                    value=7.5,
-                )
-                
-                num_inference_steps = gr.Slider(
-                    label="Inference Steps",
-                    minimum=1,
-                    maximum=25,  # Keep lower for CPU performance
-                    step=1,
-                    value=15,  # Reasonable default for CPU
-                )
+                width = gr.Slider(label="Width", minimum=256, maximum=MAX_IMAGE_SIZE, step=32, value=1024)
+                height = gr.Slider(label="Height", minimum=256, maximum=MAX_IMAGE_SIZE, step=32, value=1024)
+            guidance_scale = gr.Slider(label="Guidance scale", minimum=0.0, maximum=10.0, step=0.1, value=7.5)
+            num_inference_steps = gr.Slider(label="Number of inference steps", minimum=1, maximum=50, step=1, value=25)
         
-        gr.Examples(
-            examples=examples,
-            inputs=[prompt]
-        )
+        gr.Examples(examples=examples, inputs=[prompt])
 
-    gr.on(
-        triggers=[run_button.click, prompt.submit],
-        fn=infer,
-        inputs=[prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps],
-        outputs=[result, seed]
-    )
+    gr.on(run_button.click, infer, inputs=[prompt, negative_prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps], outputs=[result, seed])
 
 demo.queue().launch()
+