Update app.py

app.py

@@ -2,8 +2,9 @@ import torch
 from diffusers import StableDiffusion3Pipeline
 import gradio as gr
 import os
-import spaces
+from transformers import T5Tokenizer, T5ForConditionalGeneration
 from huggingface_hub import snapshot_download
+import spaces
 
 HF_TOKEN = os.getenv("HF_TOKEN")
 
@@ -27,9 +28,34 @@ else:
 pipe = StableDiffusion3Pipeline.from_pretrained(model_path, torch_dtype=torch.float16)
 pipe.to(device)
 
+tokenizer = T5Tokenizer.from_pretrained("roborovski/superprompt-v1")
+model = T5ForConditionalGeneration.from_pretrained("roborovski/superprompt-v1", device_map="auto", torch_dtype="auto")
+model.to(device)
+
 # Define the image generation function
 @spaces.GPU(duration=60)
-def generate_image(prompt, negative_prompt, num_inference_steps, height, width, guidance_scale, num_images_per_prompt):
+def generate_image(prompt, negative_prompt, num_inference_steps, height, width, guidance_scale, seed, num_images_per_prompt):
+    if seed == 0:
+        seed = random.randint(1, 2**32-1)
+        
+    if enhance_prompt:
+        transformers.set_seed(seed)
+        
+        input_text = f"Expand the following prompt to add more detail: {prompt}"
+        input_ids = tokenizer(input_text, return_tensors="pt").input_ids.to(device)
+        
+        outputs = model.generate(
+        input_ids,
+        max_new_tokens=512,
+        repetition_penalty=1.2,
+        do_sample=True,
+        temperature=0.7,
+        top_p=1,
+        top_k=50,
+    )
+        
+    generator = torch.Generator().manual_seed(seed)
+    
     output = pipe(
         prompt=prompt,
         negative_prompt=negative_prompt,
@@ -37,6 +63,7 @@ def generate_image(prompt, negative_prompt, num_inference_steps, height, width,
         height=height,
         width=width,
         guidance_scale=guidance_scale,
+        generator=generator,
         num_images_per_prompt=num_images_per_prompt
     ).images
     return output
@@ -45,6 +72,8 @@ def generate_image(prompt, negative_prompt, num_inference_steps, height, width,
 
 prompt = gr.Textbox(label="Prompt", info="Describe the image you want", placeholder="A cat...")
 
+enhance_prompt = gr.Checkbox(label="Prompt Enhancement", info="Enhance your prompt with SuperPrompt-v1", value=True)
+
 negative_prompt = gr.Textbox(label="Negative Prompt", info="Describe what you don't want in the image", placeholder="Ugly, bad anatomy...")
 
 num_inference_steps = gr.Number(label="Number of Inference Steps", precision=0, value=25)
@@ -55,11 +84,13 @@ width = gr.Slider(label="Width", info="Width of the Image", minimum=256, maximum
 
 guidance_scale = gr.Number(minimum=0.1, value=7.5, label="Guidance Scale", info="The number of denoising steps of the image. More denoising steps usually lead to a higher quality image at the cost of slower inference")
 
+seed = gr.Slider(value=42, minimum=0, maximum=2**32-1, step=1, label="Seed", info="A starting point to initiate the generation process, put 0 for a random one")
+
 num_images_per_prompt = gr.Slider(label="Number of Images to generate with the settings",minimum=1, maximum=4, step=1, value=1)
 
 interface = gr.Interface(
     fn=generate_image,
-    inputs=[prompt, negative_prompt, num_inference_steps, height, width, guidance_scale, num_images_per_prompt],
+    inputs=[prompt, enhance_prompt, negative_prompt, num_inference_steps, height, width, guidance_scale, seed, num_images_per_prompt],
     outputs=gr.Gallery(label="Generated AI Images", elem_id="gallery", show_label=False),
     title="Stable Diffusion 3 Medium",
     description="Made by <a href='https://linktr.ee/Nick088' target='_blank'>Nick088</a> \n Join https://discord.gg/osai to talk about Open Source AI"