Update app.py

app.py

@@ -214,7 +214,7 @@ def load_and_prepare_model():
     return pipe
     
 # Preload and compile both models
-models = load_and_prepare_model()
+pipe = load_and_prepare_model()
 
 MAX_SEED = np.iinfo(np.int32).max
 
@@ -238,11 +238,6 @@ def save_image(img):
     img.save(unique_name,optimize=False,compress_level=0)
     return unique_name
 
-def randomize_seed_fn(seed: int, randomize_seed: bool) -> int:
-    if randomize_seed:
-        seed = random.randint(0, MAX_SEED)
-    return seed
-
 def uploadNote(prompt,num_inference_steps,guidance_scale,timestamp):
     filename= f'tst_A_{timestamp}.txt'
     with open(filename, "w") as f:
@@ -267,22 +262,16 @@ def generate_30(
     negative_prompt: str = "",
     use_negative_prompt: bool = False,
     style_selection: str = "",
-    seed: int = 1,
     width: int = 768,
     height: int = 768,
     guidance_scale: float = 4,
     num_inference_steps: int = 125,
-    randomize_seed: bool = False,
     latent_file = gr.File(),  # Add latents file input
+    latent_file_2 = gr.File(),  # Add latents file input
     progress=gr.Progress(track_tqdm=True)  # Add progress as a keyword argument
 ):
-    #torch.backends.cudnn.benchmark = False
-    #torch.cuda.empty_cache()
-    #gc.collect()
-    global models
-    pipe = models
     ip_model = IPAdapterXL(pipe, local_folder, ip_ckpt, device)
-    seed = int(randomize_seed_fn(seed, randomize_seed))
+    seed = random.randint(0, MAX_SEED)
     generator = torch.Generator(device='cuda').manual_seed(seed)
     if latent_file is not None:  # Check if a latent file is provided
         sd_image_a = Image.open(latent_file.name)
@@ -293,8 +282,9 @@ def generate_30(
         #with torch.no_grad():
         sd_image = ip_model.generate(
                 pil_image=sd_image_a,
+                pil_image_2=sd_image_b,
                 prompt=prompt,
-                num_samples=1,
+                num_samples=samples,
                 num_inference_steps=num_inference_steps,
                 guidance_scale=guidance_scale,
                 seed=seed
@@ -312,30 +302,24 @@ def generate_30(
         image_paths = [save_image(downscale1)]     
     else:
         print('-- IMAGE REQUIRED --')
-    return image_paths, seed
+    return image_paths
 
 @spaces.GPU(duration=60)
 def generate_60(
-    prompt: str,
+    prompt: str = "",
     negative_prompt: str = "",
     use_negative_prompt: bool = False,
     style_selection: str = "",
-    seed: int = 1,
     width: int = 768,
     height: int = 768,
     guidance_scale: float = 4,
-    num_inference_steps: int = 250,
-    randomize_seed: bool = False,
+    num_inference_steps: int = 125,
     latent_file = gr.File(),  # Add latents file input
+    latent_file_2 = gr.File(),  # Add latents file input
     progress=gr.Progress(track_tqdm=True)  # Add progress as a keyword argument
 ):
-    #torch.backends.cudnn.benchmark = True
-    #torch.cuda.empty_cache()
-    #gc.collect()
-    global models
-    pipe = models
     ip_model = IPAdapterXL(pipe, local_folder, ip_ckpt, device)
-    seed = int(randomize_seed_fn(seed, randomize_seed))
+    seed = random.randint(0, MAX_SEED)
     generator = torch.Generator(device='cuda').manual_seed(seed)
     if latent_file is not None:  # Check if a latent file is provided
         sd_image_a = Image.open(latent_file.name)
@@ -346,8 +330,9 @@ def generate_60(
         #with torch.no_grad():
         sd_image = ip_model.generate(
                 pil_image=sd_image_a,
+                pil_image_2=sd_image_b,
                 prompt=prompt,
-                num_samples=1,
+                num_samples=samples,
                 num_inference_steps=num_inference_steps,
                 guidance_scale=guidance_scale,
                 seed=seed
@@ -365,30 +350,24 @@ def generate_60(
         image_paths = [save_image(downscale1)]     
     else:
         print('-- IMAGE REQUIRED --')
-    return image_paths, seed
+    return image_paths
 
 @spaces.GPU(duration=90)
 def generate_90(
-    prompt: str,
+    prompt: str = "",
     negative_prompt: str = "",
     use_negative_prompt: bool = False,
     style_selection: str = "",
-    seed: int = 1,
     width: int = 768,
     height: int = 768,
     guidance_scale: float = 4,
-    num_inference_steps: int = 250,
-    randomize_seed: bool = False,
+    num_inference_steps: int = 125,
     latent_file = gr.File(),  # Add latents file input
+    latent_file_2 = gr.File(),  # Add latents file input
     progress=gr.Progress(track_tqdm=True)  # Add progress as a keyword argument
 ):
-    #torch.backends.cudnn.benchmark = True
-    #torch.cuda.empty_cache()
-    #gc.collect()
-    global models
-    pipe = models
     ip_model = IPAdapterXL(pipe, local_folder, ip_ckpt, device)
-    seed = int(randomize_seed_fn(seed, randomize_seed))
+    seed = random.randint(0, MAX_SEED)
     generator = torch.Generator(device='cuda').manual_seed(seed)
     if latent_file is not None:  # Check if a latent file is provided
         sd_image_a = Image.open(latent_file.name)
@@ -399,8 +378,9 @@ def generate_90(
         #with torch.no_grad():
         sd_image = ip_model.generate(
                 pil_image=sd_image_a,
+                pil_image_2=sd_image_b,
                 prompt=prompt,
-                num_samples=1,
+                num_samples=samples,
                 num_inference_steps=num_inference_steps,
                 guidance_scale=guidance_scale,
                 seed=seed
@@ -418,7 +398,7 @@ def generate_90(
         image_paths = [save_image(downscale1)]     
     else:
         print('-- IMAGE REQUIRED --')
-    return image_paths, seed
+    return image_paths
 
 def load_predefined_images1():
     predefined_images1 = [
@@ -465,6 +445,7 @@ with gr.Blocks(theme=gr.themes.Origin(),css=css) as demo:
 
     with gr.Row():
         latent_file = gr.File(label="Image Prompt (Required)")
+        latent_file_2 = gr.File(label="Image Prompt 2 (Optional)")
         style_selection = gr.Radio(
             show_label=True,
             container=True,
@@ -484,10 +465,10 @@ with gr.Blocks(theme=gr.themes.Origin(),css=css) as demo:
                     value="('deformed', 'distorted', 'disfigured':1.3),'not photorealistic':1.5, 'poorly drawn', 'bad anatomy', 'wrong anatomy', 'extra limb', 'missing limb', 'floating limbs', 'poorly drawn hands', 'poorly drawn feet', 'poorly drawn face':1.3, 'out of frame', 'extra limbs', 'bad anatomy', 'bad art', 'beginner',  'distorted face','amateur'",
                     visible=True,
                 )
-        seed = gr.Slider(
+        samples = gr.Slider(
             label="Seed",
             minimum=0,
-            maximum=MAX_SEED,
+            maximum=20,
             step=1,
             value=0,
         )
@@ -547,15 +528,14 @@ with gr.Blocks(theme=gr.themes.Origin(),css=css) as demo:
             negative_prompt,
             use_negative_prompt,
             style_selection,
-            seed,
             width,
             height,
             guidance_scale,
             num_inference_steps,
-            randomize_seed, 
             latent_file,
+            latent_file_2,
         ],
-        outputs=[result, seed],
+        outputs=[result],
     )
     
     gr.on(
@@ -569,15 +549,14 @@ with gr.Blocks(theme=gr.themes.Origin(),css=css) as demo:
             negative_prompt,
             use_negative_prompt,
             style_selection,
-            seed,
             width,
             height,
             guidance_scale,
             num_inference_steps,
-            randomize_seed, 
             latent_file,
+            latent_file_2,
         ],
-        outputs=[result, seed],
+        outputs=[result],
     )
     
     gr.on(
@@ -591,15 +570,14 @@ with gr.Blocks(theme=gr.themes.Origin(),css=css) as demo:
             negative_prompt,
             use_negative_prompt,
             style_selection,
-            seed,
             width,
             height,
             guidance_scale,
             num_inference_steps,
-            randomize_seed, 
             latent_file,
+            latent_file_2,
         ],
-        outputs=[result, seed],
+        outputs=[result],
     )
 
     gr.Markdown("### REALVISXL V5.0")