Update app.py

app.py

@@ -17,20 +17,14 @@ from diffusers import AutoencoderKL, StableDiffusionXLPipeline
 from diffusers import EulerAncestralDiscreteScheduler
 #from diffusers import DPMSolverMultistepScheduler
 #from diffusers import DDIMScheduler
-#from diffusers import AutoencoderKL
 from typing import Tuple
-#from transformers import AutoTokenizer, AutoModelForCausalLM
 import paramiko
 import gc
 import time
 import datetime
 from diffusers.schedulers import AysSchedules
 
-#os.system("chmod +x ./cusparselt.sh")
-#os.system("./cusparselt.sh")
-#os.system("chmod +x ./cudnn.sh")
-#os.system("./cudnn.sh")
-#from gradio import themes
+from gradio import themes
 
 torch.backends.cuda.matmul.allow_tf32 = False
 torch.backends.cuda.matmul.allow_bf16_reduced_precision_reduction = False
@@ -48,7 +42,6 @@ FTP_USER = "ford442"
 FTP_PASS = "GoogleBez12!"
 FTP_DIR = "1ink.us/stable_diff/"  # Remote directory on FTP server
 
-#vae_url = 'https://1ink.us/files/myslrVAE_v10.safetensors'
 DESCRIPTIONXX = """
     ## REALVISXL V5.0 BF16 ⚡⚡⚡⚡
 """
@@ -113,21 +106,18 @@ def load_and_prepare_model(model_id):
     model_dtypes = {"ford442/RealVisXL_V5.0_BF16": torch.bfloat16,}
     dtype = model_dtypes.get(model_id, torch.bfloat16)  # Default to float32 if not found
     #vae = AutoencoderKL.from_pretrained("ford442/sdxl-vae-bf16", safety_checker=None)
-    vae = AutoencoderKL.from_pretrained("stabilityai/sdxl-vae")
+    #vae = AutoencoderKL.from_pretrained("stabilityai/sdxl-vae")
     #vae = AutoencoderKL.from_single_file('https://huggingface.co/ford442/sdxl-vae-bf16/mySLR/myslrVAE_v10.safetensors')
     #vae = AutoencoderKL.from_pretrained("stabilityai/sd-vae-ft-mse",use_safetensors=True)
-    #vae = AutoencoderKL.from_pretrained('ford442/Juggernaut-XI-v11-fp32',subfolder='vae')
+    vae = AutoencoderKL.from_pretrained('ford442/Juggernaut-XI-v11-fp32',subfolder='vae')
   #  vae = AutoencoderKL.from_pretrained("BeastHF/MyBack_SDXL_Juggernaut_XL_VAE/MyBack_SDXL_Juggernaut_XL_VAE_V10(version_X).safetensors",safety_checker=None).to(torch.bfloat16)
-    #  vae = AutoencoderKL.from_pretrained("ford442/sdxl-vae-bf16", safety_checker=None).to('cuda')
-    #sched = EulerAncestralDiscreteScheduler.from_pretrained('ford442/Juggernaut-XI-v11-fp32', subfolder='scheduler',beta_schedule="scaled_linear",use_karras_sigmas=True)
     #sched = EulerAncestralDiscreteScheduler.from_pretrained("SG161222/RealVisXL_V5.0", subfolder='scheduler',beta_schedule="scaled_linear", steps_offset=1,timestep_spacing="trailing"))
     sched = EulerAncestralDiscreteScheduler.from_pretrained("SG161222/RealVisXL_V5.0", subfolder='scheduler', steps_offset=1,timestep_spacing="trailing")
-    #sched = EulerAncestralDiscreteScheduler(timestep_spacing="trailing",steps_offset=1)
     #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 = EulerAncestralDiscreteScheduler.from_config('ford442/RealVisXL_V5.0_BF16', beta_schedule="scaled_linear", beta_start=0.00085, beta_end=0.012, steps_offset=1)
-    pipeX = StableDiffusionXLPipeline.from_pretrained("SG161222/RealVisXL_V5.0")
-    
+    pipeX = StableDiffusionXLPipeline.from_pretrained("SG161222/RealVisXL_V5.0",use_safetensors=False)
     #pipeX = StableDiffusionXLPipeline.from_pretrained("ford442/Juggernaut-XI-v11-fp32",use_safetensors=True)
+    
     pipe = StableDiffusionXLPipeline.from_pretrained(
         'ford442/RealVisXL_V5.0_BF16',
     #    'ford442/Juggernaut-XI-v11-fp32',
@@ -146,18 +136,15 @@ def load_and_prepare_model(model_id):
       #  scheduler = EulerAncestralDiscreteScheduler.from_config(pipeX.scheduler.config, beta_schedule="scaled_linear", beta_start=0.00085, beta_end=0.012, steps_offset=1)
         #scheduler=EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config, beta_schedule="scaled_linear", beta_start=0.00085, beta_end=0.012, steps_offset =1)
     )
+    
     #sched = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config, beta_schedule="scaled_linear",use_karras_sigmas=True, algorithm_type="dpmsolver++")
     #pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config, beta_schedule="scaled_linear", beta_start=0.00085, beta_end=0.012, steps_offset=1)
-    #pipe.to('cuda')
+      # for set timestep  pipeline.scheduler = DPMSolverMultistepScheduler.from_config(pipeline.scheduler.config, algorithm_type="sde-dpmsolver++")
 
     pipe.scheduler = sched
-    pipe.vae=vae
-    pipe.unet=pipeX.unet
-   # pipe.scheduler=EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config)
-    #pipe.to(dtype=torch.bfloat16)
-    #pipe.unet = pipeX.unet
+    pipe.vae=vae.to(torch.bfloat16)
+    pipe.unet=pipeX.unet.to(torch.bfloat16)
     #pipe.scheduler=EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config, beta_schedule="scaled_linear", beta_start=0.00085, beta_end=0.012, steps_offset=1)
-    #pipe.unet.to(torch.bfloat16)
 
     pipe.to(device)
     pipe.to(torch.bfloat16)
@@ -165,7 +152,6 @@ def load_and_prepare_model(model_id):
     pipe.unet.set_default_attn_processor()
     pipe.vae.set_default_attn_processor()
     
-    #pipe.to(torch.bfloat16)
     print(f'Pipeline: ')
     print(f'_optional_components: {pipe._optional_components}')
     print(f'watermark: {pipe.watermark}')