pre load pipeline pipe_class

check_app.py

@@ -22,7 +22,7 @@ from datetime import datetime
 MAX_SEED = np.iinfo(np.int32).max
 MAX_IMAGE_SIZE = 1024
     
-
+pipe = {}
     
 class ProgressPipeline(DiffusionPipeline):
     def __init__(self, original_pipeline):
@@ -180,24 +180,26 @@ def create_pipeline_logic(prompt_text, model_name, negative_prompt="",  seed=42,
     progress = gr.Progress(track_tqdm=False)
     config = MODEL_CONFIGS[model_name]
     pipe_class = config["pipeline_class"]
-    pipe = None
-    b_pipe = AutoPipelineForText2Image.from_pretrained(
-        config["repo_id"],
-        #variant="fp16",
-        #cache_dir=config["cache_dir"],
-        torch_dtype=torch.bfloat16
-    ).to("cuda")
-    pipe_signature = signature(b_pipe)
-    # Check for the presence of "callback_on_step_end" in the signature
-    has_callback_on_step_end = "callback_on_step_end" in pipe_signature.parameters
-    if not has_callback_on_step_end:
-        pipe = ProgressPipeline(b_pipe)
-        print("ProgressPipeline specal")
-    else:
-        pipe = b_pipe
+    global pipe
+    if ["pipeline_class"] != pipe_class
+        pipe["pipeline_class"] = pipe_class
+        b_pipe = AutoPipelineForText2Image.from_pretrained(
+            config["repo_id"],
+            #variant="fp16",
+            #cache_dir=config["cache_dir"],
+            torch_dtype=torch.bfloat16
+        ).to("cuda")
+        pipe_signature = signature(b_pipe)
+        # Check for the presence of "callback_on_step_end" in the signature
+        has_callback_on_step_end = "callback_on_step_end" in pipe_signature.parameters
+        if not has_callback_on_step_end:
+            pipe["pipeline"] = ProgressPipeline(b_pipe)
+            print("ProgressPipeline specal")
+        else:
+            pipe["pipeline"] = b_pipe
         
     gen_seed,image, images = generate_image_with_progress(
-        model_name,pipe, prompt_text, num_steps=num_inference_steps, guidance_scale=guidance_scale, seed=seed,negative_prompt = negative_prompt,  randomize_seed = randomize_seed, width = width, height = height, progress=progress
+        model_name,pipe["pipeline"], prompt_text, num_steps=num_inference_steps, guidance_scale=guidance_scale, seed=seed,negative_prompt = negative_prompt,  randomize_seed = randomize_seed, width = width, height = height, progress=progress
     )
     return f"Seed: {gen_seed}", image, images
 def main():