Update app.py

app.py

@@ -1,7 +1,7 @@
 from diffusers import StableDiffusionPipeline
 import gc
 
-pipe = StableDiffusionPipeline.from_pretrained("prompthero/openjourney-v4").to("cpu")
+pipe = StableDiffusionPipeline.from_pretrained("CompVis/stable-diffusion-v1-4").to("cpu")
 text_encoder = pipe.text_encoder
 text_encoder.eval()
 unet = pipe.unet
@@ -43,7 +43,6 @@ def convert_encoder(text_encoder:torch.nn.Module,ir_path:Path):
         del ov_model
         cleanup_cache()
         print(f"Text Encoder successfully converted to TR and saved to {ir_path}")
-
 if not text_encoder_path.exists():
     convert_encoder(text_encoder,text_encoder_path)
 else:
@@ -425,8 +424,8 @@ class OVStableDiffusionPipeline(DiffusionPipeline):
 
         return timesteps, num_inference_steps - t_start 
 
-core=pv.Core()
 
+core=pv.Core()
 import ipywidgets as widgets
 device=widgets.Dropdown(
     options=core.available_devices+["AUTO"],
@@ -435,11 +434,12 @@ device=widgets.Dropdown(
     disabled=False,
 )
 device
-
-
 text_enc=core.compile_model(text_encoder_path,device.value)
-
 unet_model=core.compile_model(unet_path,device.value)
+
+pv_config={"INFERENCE_PRECISION_HINT":"f32"}if device.value !="CPU" else {}
+vae_decoder=core.compile_model(VAE_DECODER_PATH,device.value,pv_config)
+vae_encoder=core.compile_model(VAE_ENCODER_PATH,device.value,pv_config)
 from transformers import CLIPTokenizer
 from diffusers.schedulers import LMSDiscreteScheduler
 lms=LMSDiscreteScheduler(
@@ -455,6 +455,13 @@ pv_pipe=OVStableDiffusionPipeline(
     vae_encoder=vae_encoder,
     vae_decoder=vae_decoder,
     scheduler=lms)
+from ipywidgets import widgets
+sample_text=("A Dog wearing golden rich mens necklace")
+text_prompt=widgets.Text(value=sample_text,description="A Dog wearing golden rich mens necklace ")
+num_steps=widgets.IntSlider(min=1,max=50,value=20,description="steps:")
+seed=widgets.IntSlider(min=0,max=10000000,description="seed:",value=54)
+widgets.VBox([text_prompt,seed,num_steps])
+result=pv_pipe(text_prompt.value,num_inference_steps=num_steps.value,seed=seed.value)
 
 import gradio as gr
 
@@ -479,5 +486,4 @@ with gr.Blocks() as demo:
 try:
     demo.queue().launch(debug=True)
 except Exception:
-    demo.queue().launch(share=True,debug=True)
-    
+    demo.queue().launch(share=True,debug=True)