Update app.py

app.py

@@ -30,7 +30,7 @@ from pipeline_stable_diffusion_xl_instantid_img2img import StableDiffusionXLInst
 from controlnet_aux import ZoeDetector
 
 from compel import Compel, ReturnedEmbeddingsType
-#import spaces
+import spaces
 
 #from gradio_imageslider import ImageSlider
 
@@ -123,8 +123,6 @@ pipe.load_ip_adapter_instantid(face_adapter)
 pipe.set_ip_adapter_scale(0.8)
 zoe = ZoeDetector.from_pretrained("lllyasviel/Annotators")
 zoe.to(device)
-
-original_pipe = copy.deepcopy(pipe)
 pipe.to(device)
 
 last_lora = ""
@@ -204,32 +202,7 @@ def merge_incompatible_lora(full_path_lora, lora_scale):
                 )
                 del weights_sd
                 del lora_model
-#@spaces.GPU
-def generate_image(prompt, negative, face_emb, face_image, image_strength, images, guidance_scale, face_strength, depth_control_scale):
-    print("Processing prompt...")
-    conditioning, pooled = compel(prompt)
-    if(negative):
-        negative_conditioning, negative_pooled = compel(negative)
-    else:
-        negative_conditioning, negative_pooled = None, None
-    print("Processing image...")
-    image = pipe(
-        prompt_embeds=conditioning,
-        pooled_prompt_embeds=pooled,
-        negative_prompt_embeds=negative_conditioning,
-        negative_pooled_prompt_embeds=negative_pooled,
-        width=1024,
-        height=1024,
-        image_embeds=face_emb,
-        image=face_image,
-        strength=1-image_strength,
-        control_image=images,
-        num_inference_steps=20,
-        guidance_scale = guidance_scale,
-        controlnet_conditioning_scale=[face_strength, depth_control_scale],
-    ).images[0]
-    return image
-    
+@spaces.GPU    
 def run_lora(face_image, prompt, negative, lora_scale, selected_state, face_strength, image_strength, guidance_scale, depth_control_scale, sdxl_loras, progress=gr.Progress(track_tqdm=True)):
     global last_lora, last_merged, last_fused, pipe
 
@@ -304,8 +277,31 @@ def run_lora(face_image, prompt, negative, lora_scale, selected_state, face_stre
                 pipe.unload_textual_inversion()
                 pipe.load_textual_inversion(state_dict_embedding["text_encoders_0"], token=["<s0>", "<s1>"], text_encoder=pipe.text_encoder, tokenizer=pipe.tokenizer)
                 pipe.load_textual_inversion(state_dict_embedding["text_encoders_1"], token=["<s0>", "<s1>"], text_encoder=pipe.text_encoder_2, tokenizer=pipe.tokenizer_2)
-
-    image = generate_image(prompt, negative, face_emb, face_image, image_strength, images, guidance_scale, face_strength, depth_control_scale)
+    
+    print("Processing prompt...")
+    conditioning, pooled = compel(prompt)
+    if(negative):
+        negative_conditioning, negative_pooled = compel(negative)
+    else:
+        negative_conditioning, negative_pooled = None, None
+    print("Processing image...")
+    
+    image = pipe(
+        prompt_embeds=conditioning,
+        pooled_prompt_embeds=pooled,
+        negative_prompt_embeds=negative_conditioning,
+        negative_pooled_prompt_embeds=negative_pooled,
+        width=1024,
+        height=1024,
+        image_embeds=face_emb,
+        image=face_image,
+        strength=1-image_strength,
+        control_image=images,
+        num_inference_steps=20,
+        guidance_scale = guidance_scale,
+        controlnet_conditioning_scale=[face_strength, depth_control_scale],
+    ).images[0]
+    
     last_lora = repo_name
     return image, gr.update(visible=True)