Update app.py

app.py

@@ -83,20 +83,15 @@ os.putenv("HF_HUB_ENABLE_HF_TRANSFER","1")
 
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
-text_encoder=CLIPTextModel.from_pretrained('ford442/RealVisXL_V5.0_BF16', low_cpu_mem_usage=False, subfolder='text_encoder',token=True)#.to(device=device, dtype=torch.bfloat16)
-text_encoder_2=CLIPTextModelWithProjection.from_pretrained('ford442/RealVisXL_V5.0_BF16', low_cpu_mem_usage=False, subfolder='text_encoder_2',token=True)#.to(device=device, dtype=torch.bfloat16)
-tokenizer_1=CLIPTokenizer.from_pretrained('ford442/RealVisXL_V5.0_BF16', low_cpu_mem_usage=False, subfolder='tokenizer',token=True)
-tokenizer_2=CLIPTokenizer.from_pretrained('ford442/RealVisXL_V5.0_BF16', low_cpu_mem_usage=False, subfolder='tokenizer_2',token=True)
-scheduler=EulerAncestralDiscreteScheduler.from_pretrained('ford442/RealVisXL_V5.0_BF16', low_cpu_mem_usage=False, subfolder='scheduler',token=True)
+text_encoder = CLIPTextModel.from_pretrained('ford442/RealVisXL_V5.0_BF16', low_cpu_mem_usage=False, subfolder='text_encoder', token=True)#.to(device=device, dtype=torch.bfloat16)
+text_encoder_2 = CLIPTextModelWithProjection.from_pretrained('ford442/RealVisXL_V5.0_BF16', low_cpu_mem_usage=False, subfolder='text_encoder_2',token=True)#.to(device=device, dtype=torch.bfloat16)
+tokenizer_1 = CLIPTokenizer.from_pretrained('ford442/RealVisXL_V5.0_BF16', low_cpu_mem_usage=False, subfolder='tokenizer', token=True)
+tokenizer_2 = CLIPTokenizer.from_pretrained('ford442/RealVisXL_V5.0_BF16', low_cpu_mem_usage=False, subfolder='tokenizer_2', token=True)
+scheduler = EulerAncestralDiscreteScheduler.from_pretrained('ford442/RealVisXL_V5.0_BF16', low_cpu_mem_usage=False, subfolder='scheduler', token=True)
 vaeXL = AutoencoderKL.from_pretrained("stabilityai/sdxl-vae", low_cpu_mem_usage=False, safety_checker=None, use_safetensors=False, torch_dtype=torch.float32, token=True) #.to(device).to(torch.bfloat16) #.to(device=device, dtype=torch.bfloat16)
-UNet2DConditionModel
+unet = UNet2DConditionModel.from_pretrained("stabilityai/sdxl-vae", low_cpu_mem_usage=False, subfolder='unet', token=True)
+
 def load_and_prepare_model():
-    #vaeRV = AutoencoderKL.from_pretrained("SG161222/RealVisXL_V5.0", subfolder='vae', safety_checker=None, use_safetensors=True, token=True)
-    #vaeXL = AutoencoderKL.from_pretrained("stabilityai/sdxl-vae", safety_checker=None, use_safetensors=False, low_cpu_mem_usage=False, torch_dtype=torch.float32, token=True) #.to(device).to(torch.bfloat16) #.to(device=device, dtype=torch.bfloat16)
-    #sched = EulerAncestralDiscreteScheduler.from_pretrained('ford442/RealVisXL_V5.0_BF16', subfolder='scheduler',beta_schedule="scaled_linear", beta_start=0.00085, beta_end=0.012, steps_offset=1,use_karras_sigmas=True)
-    #sched = DPMSolverSDEScheduler.from_pretrained('ford442/RealVisXL_V5.0_BF16', subfolder='scheduler')
-    #sched = EulerAncestralDiscreteScheduler.from_pretrained('ford442/RealVisXL_V5.0_BF16', subfolder='scheduler',beta_schedule="scaled_linear", token=True) #, beta_start=0.00085, beta_end=0.012, steps_offset=1,use_karras_sigmas=True, token=True)
-    #sched = EulerAncestralDiscreteScheduler.from_pretrained('ford442/RealVisXL_V5.0_BF16', subfolder='scheduler',beta_schedule="scaled_linear")
     pipe = StableDiffusionXLPipeline.from_pretrained(
         'ford442/RealVisXL_V5.0_BF16',
         #torch_dtype=torch.bfloat16,
@@ -185,7 +180,7 @@ def generate_30(
     guidance_scale: float = 4,
     num_inference_steps: int = 125,
     use_resolution_binning: bool = True, 
-    progress=gr.Progress(track_tqdm=True)  # Add progress as a keyword argument
+    progress=gr.Progress(track_tqdm=True)
 ):
     seed = random.randint(0, MAX_SEED)
     generator = torch.Generator(device='cuda').manual_seed(seed)
@@ -227,7 +222,7 @@ def generate_60(
     guidance_scale: float = 4,
     num_inference_steps: int = 125,
     use_resolution_binning: bool = True, 
-    progress=gr.Progress(track_tqdm=True)  # Add progress as a keyword argument
+    progress=gr.Progress(track_tqdm=True)
 ):
     seed = random.randint(0, MAX_SEED)
     generator = torch.Generator(device='cuda').manual_seed(seed)
@@ -269,7 +264,7 @@ def generate_90(
     guidance_scale: float = 4,
     num_inference_steps: int = 125,
     use_resolution_binning: bool = True, 
-    progress=gr.Progress(track_tqdm=True)  # Add progress as a keyword argument
+    progress=gr.Progress(track_tqdm=True)
 ):
     seed = random.randint(0, MAX_SEED)
     generator = torch.Generator(device='cuda').manual_seed(seed)
@@ -341,6 +336,7 @@ with gr.Blocks(theme=gr.themes.Origin(),css=css) as demo:
         run_button_30 = gr.Button("Run 30 Seconds", scale=0)
         run_button_60 = gr.Button("Run 60 Seconds", scale=0)
         run_button_90 = gr.Button("Run 90 Seconds", scale=0)
+        
     result = gr.Gallery(label="Result", columns=1, show_label=False) 
 
     with gr.Row():