Update app.py

app.py

@@ -3,7 +3,19 @@ import json
 import logging
 import torch
 from PIL import Image
-from diffusers import DiffusionPipeline, EulerDiscreteScheduler, DPMSolverMultistepScheduler
+from diffusers import (
+    DiffusionPipeline, 
+    EulerDiscreteScheduler, 
+    DPMSolverMultistepScheduler,
+    DPMSolverSinglestepScheduler,
+    KDPM2DiscreteScheduler,
+    KDPM2AncestralDiscreteScheduler,
+    EulerAncestralDiscreteScheduler,
+    HeunDiscreteScheduler,
+    LMSDiscreteScheduler,
+    DEISMultistepScheduler,
+    UniPCMultistepScheduler
+)
 import spaces
 
 # Load LoRAs from JSON file
@@ -27,7 +39,7 @@ def update_selection(evt: gr.SelectData):
     )
 
 @spaces.GPU
-def run_lora(prompt, negative_prompt, cfg_scale, steps, selected_index, scheduler):
+def run_lora(prompt, negative_prompt, cfg_scale, steps, selected_index, scheduler, seed, width, height, lora_scale):
     if selected_index is None:
         raise gr.Error("You must select a LoRA before proceeding.")
 
@@ -39,10 +51,44 @@ def run_lora(prompt, negative_prompt, cfg_scale, steps, selected_index, schedule
     pipe.load_lora_weights(lora_path)
 
     # Set scheduler
-    if scheduler == "Euler":
-        pipe.scheduler = EulerDiscreteScheduler.from_config(pipe.scheduler.config)
-    elif scheduler == "DPM++ 2M":
-        pipe.scheduler = DPMSolverMultistepScheduler.from_config(pipe.scheduler.config)
+    scheduler_config = pipe.scheduler.config
+    if scheduler == "DPM++ 2M":
+        pipe.scheduler = DPMSolverMultistepScheduler.from_config(scheduler_config)
+    elif scheduler == "DPM++ 2M Karras":
+        pipe.scheduler = DPMSolverMultistepScheduler.from_config(scheduler_config, use_karras_sigmas=True)
+    elif scheduler == "DPM++ 2M SDE":
+        pipe.scheduler = DPMSolverMultistepScheduler.from_config(scheduler_config, algorithm_type="sde-dpmsolver++")
+    elif scheduler == "DPM++ 2M SDE Karras":
+        pipe.scheduler = DPMSolverMultistepScheduler.from_config(scheduler_config, use_karras_sigmas=True, algorithm_type="sde-dpmsolver++")
+    elif scheduler == "DPM++ SDE":
+        pipe.scheduler = DPMSolverSinglestepScheduler.from_config(scheduler_config)
+    elif scheduler == "DPM++ SDE Karras":
+        pipe.scheduler = DPMSolverSinglestepScheduler.from_config(scheduler_config, use_karras_sigmas=True)
+    elif scheduler == "DPM2":
+        pipe.scheduler = KDPM2DiscreteScheduler.from_config(scheduler_config)
+    elif scheduler == "DPM2 Karras":
+        pipe.scheduler = KDPM2DiscreteScheduler.from_config(scheduler_config, use_karras_sigmas=True)
+    elif scheduler == "DPM2 a":
+        pipe.scheduler = KDPM2AncestralDiscreteScheduler.from_config(scheduler_config)
+    elif scheduler == "DPM2 a Karras":
+        pipe.scheduler = KDPM2AncestralDiscreteScheduler.from_config(scheduler_config, use_karras_sigmas=True)
+    elif scheduler == "Euler":
+        pipe.scheduler = EulerDiscreteScheduler.from_config(scheduler_config)
+    elif scheduler == "Euler a":
+        pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(scheduler_config)
+    elif scheduler == "Heun":
+        pipe.scheduler = HeunDiscreteScheduler.from_config(scheduler_config)
+    elif scheduler == "LMS":
+        pipe.scheduler = LMSDiscreteScheduler.from_config(scheduler_config)
+    elif scheduler == "LMS Karras":
+        pipe.scheduler = LMSDiscreteScheduler.from_config(scheduler_config, use_karras_sigmas=True)
+    elif scheduler == "DEIS":
+        pipe.scheduler = DEISMultistepScheduler.from_config(scheduler_config)
+    elif scheduler == "UniPC":
+        pipe.scheduler = UniPCMultistepScheduler.from_config(scheduler_config)
+
+    # Set random seed for reproducibility
+    generator = torch.Generator(device="cuda").manual_seed(seed)
 
     # Generate image
     image = pipe(
@@ -50,6 +96,10 @@ def run_lora(prompt, negative_prompt, cfg_scale, steps, selected_index, schedule
         negative_prompt=negative_prompt,
         num_inference_steps=steps,
         guidance_scale=cfg_scale,
+        width=width,
+        height=height,
+        generator=generator,
+        cross_attention_kwargs={"scale": lora_scale},
     ).images[0]
 
     # Unload LoRA weights
@@ -57,8 +107,8 @@ def run_lora(prompt, negative_prompt, cfg_scale, steps, selected_index, schedule
 
     return image
 
-with gr.Blocks(css="custom.css") as app:
-    gr.Markdown("# artificialguybr LoRA portfolio")
+with gr.Blocks(theme=gr.themes.Soft()) as app:
+    gr.Markdown("# artificialguybr LoRA Portfolio")
     gr.Markdown(
         "### This is my portfolio. Follow me on Twitter [@artificialguybr](https://twitter.com/artificialguybr).\n"
         "**Note**: Generation quality may vary. For best results, adjust the parameters.\n"
@@ -68,33 +118,53 @@ with gr.Blocks(css="custom.css") as app:
     selected_index = gr.State(None)
 
     with gr.Row():
-        gallery = gr.Gallery(
-            [(item["image"], item["title"]) for item in loras],
-            label="LoRA Gallery",
-            allow_preview=False,
-            columns=3
-        )
+        with gr.Column(scale=2):
+            result = gr.Image(label="Generated Image", height=768)
+            generate_button = gr.Button("Generate", variant="primary")
+
+        with gr.Column(scale=1):
+            gallery = gr.Gallery(
+                [(item["image"], item["title"]) for item in loras],
+                label="LoRA Gallery",
+                allow_preview=False,
+                columns=2
+            )
 
+    with gr.Row():
         with gr.Column():
             prompt_title = gr.Markdown("### Click on a LoRA in the gallery to select it")
             selected_info = gr.Markdown("")
             prompt = gr.Textbox(label="Prompt", lines=3, placeholder="Type a prompt after selecting a LoRA")
             negative_prompt = gr.Textbox(label="Negative Prompt", lines=2, value="low quality, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry")
-            
+
+        with gr.Column():
             with gr.Row():
                 cfg_scale = gr.Slider(label="CFG Scale", minimum=1, maximum=20, step=0.5, value=7.5)
                 steps = gr.Slider(label="Steps", minimum=1, maximum=100, step=1, value=30)
             
-            scheduler = gr.Dropdown(label="Scheduler", choices=["Euler", "DPM++ 2M"], value="Euler")
+            with gr.Row():
+                width = gr.Slider(label="Width", minimum=256, maximum=1536, step=64, value=1024)
+                height = gr.Slider(label="Height", minimum=256, maximum=1536, step=64, value=1024)
+            
+            with gr.Row():
+                seed = gr.Slider(label="Seed", minimum=0, maximum=2**32-1, step=1, value=0, randomize=True)
+                lora_scale = gr.Slider(label="LoRA Scale", minimum=0, maximum=1, step=0.01, value=0.75)
             
-            generate_button = gr.Button("Generate")
-            result = gr.Image(label="Generated Image")
+            scheduler = gr.Dropdown(
+                label="Scheduler", 
+                choices=[
+                    "DPM++ 2M", "DPM++ 2M Karras", "DPM++ 2M SDE", "DPM++ 2M SDE Karras",
+                    "DPM++ SDE", "DPM++ SDE Karras", "DPM2", "DPM2 Karras", "DPM2 a", "DPM2 a Karras",
+                    "Euler", "Euler a", "Heun", "LMS", "LMS Karras", "DEIS", "UniPC"
+                ],
+                value="Euler"
+            )
 
     gallery.select(update_selection, outputs=[prompt, selected_info, selected_index])
     
     generate_button.click(
         fn=run_lora,
-        inputs=[prompt, negative_prompt, cfg_scale, steps, selected_index, scheduler],
+        inputs=[prompt, negative_prompt, cfg_scale, steps, selected_index, scheduler, seed, width, height, lora_scale],
         outputs=[result]
     )