Update app.py

app.py

@@ -2,19 +2,31 @@ import gradio as gr
 import random
 import os
 
-# model = gr.load("models/prithivMLmods/SD3.5-Turbo-Realism-2.0-LoRA") # <-- This model is also working
+
+default_negative_prompt = (
+    "Extra limbs, Extra fingers or toes, Disfigured face, Distorted hands, Mutated body parts, "
+    "Missing limbs, Asymmetrical features, Blurry face, Poor anatomy, Incorrect proportions, Crooked eyes, "
+    "Deformed hands or fingers, Double face, Unrealistic skin texture, Overly smooth skin, Poor lighting, "
+    "Cartoonish appearance, Plastic look, Grainy, Unnatural expressions, Crossed eyes, Mutated clothing, "
+    "Artifacts, Uncanny valley, Doll-like features, Bad symmetry, Uneven skin tones, Extra teeth, "
+    "Unrealistic hair texture, Dark shadows on face, Overexposed face, Cluttered background, Text, watermark, "
+    "or signature, Over-processed, Low quality, Blurry, Low resolution, Pixelated, Oversaturated, Too dark, Overexposed, Poor lighting, "
+    "Unclear, Text or watermark, Distorted faces, Extra limbs or fingers, Disfigured, Grainy, Overly stylized, Cartoonish, "
+    "Unrealistic anatomy, Bad proportions, Unrealistic colors, Artificial look, Background noise, Unwanted objects, Repetitive patterns, Artifacting, Abstract shapes, Out of focus"
+)
+
 model = gr.load("models/Purz/face-projection")
 
-def generate_image(text, seed, width, height, guidance_scale, num_inference_steps):
+def generate_image(text, seed, width, height, guidance_scale, num_inference_steps, negative_prompt):
     if seed is not None:
         random.seed(seed)
 
     if text in [example[0] for example in examples]:
         print(f"Using example: {text}")
 
-    result_image = model(text)
+    result_image = model(text, negative_prompt=negative_prompt)
     
-    print(f"Width: {width}, Height: {height}, Guidance Scale: {guidance_scale}, Inference Steps: {num_inference_steps}")
+    print(f"Width: {width}, Height: {height}, Guidance Scale: {guidance_scale}, Inference Steps: {num_inference_steps}, Negative Prompt: {negative_prompt}")
     
     return result_image
 
@@ -43,6 +55,15 @@ interface = gr.Interface(
         gr.Slider(label="Height", minimum=512, maximum=2048, step=64, value=1024),
         gr.Slider(label="Guidance Scale", minimum=0.1, maximum=20.0, step=0.1, value=3.0),
         gr.Slider(label="Number of inference steps", minimum=1, maximum=40, step=1, value=28),
+        gr.Dropdown(
+            label="Negative Prompt",
+            choices=[
+                "Default Negative Prompt",
+                "None",
+            ],
+            value="Default Negative Prompt",
+            description="Choose a negative prompt to apply to the image generation."
+        ),
     ],
     outputs=gr.Image(label="Generated Image"),
     examples=examples,
@@ -50,4 +71,17 @@ interface = gr.Interface(
     description="Sorry for the inconvenience. The model is currently running on the CPU, which might affect performance. We appreciate your understanding.",
 )
 
-interface.launch()
+def get_negative_prompt(negative_prompt_choice):
+    if negative_prompt_choice == "Default Negative Prompt":
+        return default_negative_prompt
+    else:
+        return ""
+
+
+def generate_image_with_negative_prompt(text, seed, width, height, guidance_scale, num_inference_steps, negative_prompt_choice):
+    negative_prompt = get_negative_prompt(negative_prompt_choice)
+    return generate_image(text, seed, width, height, guidance_scale, num_inference_steps, negative_prompt)
+
+interface.fn = generate_image_with_negative_prompt
+
+interface.launch()