Create app.py

app.py

@@ -0,0 +1,44 @@
+from huggingface_hub import from_pretrained_keras
+from keras_cv import models
+import gradio as gr
+import tensorflow as tf 
+
+# load keras model
+resolution = 512
+dreambooth_model = models.StableDiffusion(
+        img_width=resolution, img_height=resolution, jit_compile=True, 
+    )
+loaded_diffusion_model = from_pretrained_keras("keras-dreambooth/dreambooth_diffusion_hokusai")
+dreambooth_model._diffusion_model = loaded_diffusion_model
+
+
+# generate images
+def inference(prompt, negative_prompt, num_imgs_to_gen, num_steps, guidance_scale):
+    generated_images = dreambooth_model.text_to_image(
+        prompt,
+        negative_prompt=negative_prompt,
+        batch_size=num_imgs_to_gen,
+        num_steps=num_steps,
+        unconditional_guidance_scale=guidance_scale,
+    )
+    return generated_images 
+    
+    
+
+# pass function, input type for prompt, the output for multiple images
+gr.Interface(
+    inference, [
+        gr.Textbox(label="Positive Prompt", value="a painting image in hks## style"),
+        gr.Textbox(label="Negative Prompt", value="bad anatomy, soft blurry"),
+        gr.Slider(label='Number of gen image', minimum=1, maximum=4, value=2, step=1),
+        gr.Slider(label="Inference Steps",value=100),
+        gr.Number(label='Guidance scale', value=7.5),
+    ], [
+        gr.Gallery(show_label=False),
+    ],
+    title="Keras Dreambooth - Hokusai artist 🏞",
+    description = "This model has been fine-tuned to learn the concept of Hokusai artist. To use this demo, you should have {hks## style} in the input",
+    examples = [["a painting image of a fishing village under a cherry blossom forest at sunset in hks## style, ultra realistic, 4k, 8k", "bad anatomy, soft blurry", 4, 100, 15],
+               ["a beautiful and highly detailed oil painting of a lost valley in the mountains in hks## style, running river, intricate details, 8k, sharp focus, hyper realism", "(bad anatomy), (blurry), grain", 4, 100, 15]],
+    cache_examples=True
+    ).queue().launch(debug=True, share=True)