Create app.py

app.py

@@ -0,0 +1,171 @@
+import gradio as gr
+import numpy as np
+import random
+import spaces
+import torch
+from diffusers import DiffusionPipeline, FlowMatchEulerDiscreteScheduler
+from transformers import CLIPTextModel, CLIPTokenizer, T5EncoderModel, T5TokenizerFast
+
+# Initialize model and settings
+dtype = torch.bfloat16
+device = "cuda" if torch.cuda.is_available() else "cpu"
+pipe = DiffusionPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16).to(device)
+
+# Constants
+MAX_SEED = np.iinfo(np.int32).max
+MAX_IMAGE_SIZE = 2048
+
+@spaces.GPU(duration=190)
+def infer(prompt, seed=42, randomize_seed=False, width=1024, height=1024, guidance_scale=5.0, num_inference_steps=28, output_format="png", progress=gr.Progress(track_tqdm=True)):
+    if randomize_seed:
+        seed = random.randint(0, MAX_SEED)
+    generator = torch.Generator(device=device).manual_seed(seed)
+    image = pipe(
+        prompt=prompt, 
+        width=width, 
+        height=height, 
+        num_inference_steps=num_inference_steps, 
+        generator=generator, 
+        guidance_scale=guidance_scale
+    ).images[0]
+
+    # Convert image to desired format
+    if output_format.lower() != "png":
+        image = image.convert(output_format.upper())
+    
+    return image, seed
+
+examples = [
+    "a tiny astronaut hatching from an egg on the moon",
+    "a cat holding a sign that says hello world",
+    "an anime illustration of a wiener schnitzel",
+]
+
+css = """
+#col-container {
+    margin: 0 auto;
+    max-width: 800px;
+    padding: 20px;
+    background-color: #f9f9f9;
+    border-radius: 10px;
+    box-shadow: 0px 0px 20px rgba(0,0,0,0.1);
+}
+
+#title {
+    font-family: 'Arial', sans-serif;
+    color: #333;
+    text-align: center;
+    margin-bottom: 20px;
+}
+
+#advanced-settings {
+    background-color: #f1f1f1;
+    border-radius: 8px;
+    padding: 10px;
+}
+
+#output-container {
+    text-align: center;
+    margin-top: 20px;
+}
+"""
+
+with gr.Blocks(css=css) as demo:
+    with gr.Column(elem_id="col-container"):
+        
+        # Title and Description
+        gr.Markdown(f"""
+        <h1 id="title">FLUX.1 [dev] - Advanced Text-to-Image Generator</h1>
+        <p style="text-align:center; color:#555;">
+        Experience the power of a 12B param rectified flow transformer. Customize your prompts and settings to generate unique images every time.
+        </p>
+        """)
+        
+        with gr.Row():
+            prompt = gr.Textbox(
+                label="Prompt",
+                show_label=False,
+                max_lines=1,
+                placeholder="Enter your creative prompt...",
+                container=False,
+                interactive=True
+            )
+        
+        run_button = gr.Button("Generate", elem_id="generate-button")
+
+        # Output image and settings
+        with gr.Row(elem_id="output-container"):
+            result = gr.Image(label="Generated Image", show_label=False).style(height=400)
+            output_format = gr.Radio(
+                label="Output Format",
+                choices=["png", "jpeg", "bmp"],
+                value="png",
+                interactive=True
+            )
+
+        # Advanced settings
+        with gr.Accordion("Advanced Settings", open=False, elem_id="advanced-settings"):
+            seed = gr.Slider(
+                label="Seed",
+                minimum=0,
+                maximum=MAX_SEED,
+                step=1,
+                value=0,
+                interactive=True
+            )
+            randomize_seed = gr.Checkbox(label="Randomize Seed", value=True, interactive=True)
+
+            with gr.Row():
+                width = gr.Slider(
+                    label="Width",
+                    minimum=256,
+                    maximum=MAX_IMAGE_SIZE,
+                    step=32,
+                    value=1024,
+                    interactive=True
+                )
+                height = gr.Slider(
+                    label="Height",
+                    minimum=256,
+                    maximum=MAX_IMAGE_SIZE,
+                    step=32,
+                    value=1024,
+                    interactive=True
+                )
+
+            with gr.Row():
+                guidance_scale = gr.Slider(
+                    label="Guidance Scale",
+                    minimum=1,
+                    maximum=15,
+                    step=0.1,
+                    value=5.0,
+                    interactive=True
+                )
+                num_inference_steps = gr.Slider(
+                    label="Number of Inference Steps",
+                    minimum=1,
+                    maximum=50,
+                    step=1,
+                    value=28,
+                    interactive=True
+                )
+
+        # Interactive Examples
+        gr.Examples(
+            examples=examples,
+            fn=infer,
+            inputs=[prompt],
+            outputs=[result, seed],
+            cache_examples="lazy",
+            label="Try these examples:"
+        )
+
+        # Link button to trigger inference
+        run_button.click(
+            fn=infer,
+            inputs=[prompt, seed, randomize_seed, width, height, guidance_scale, num_inference_steps, output_format],
+            outputs=[result, seed]
+        )
+
+demo.launch()