chore: Update inpainting pipeline configuration and parameters

Former-commit-id: 9ccf8e9ea2ff6e8b5d91537a79af334e8ada1c12 [formerly 8312cab660b7e3f1ac5bf3ecf8f3b9aec6883906]
Former-commit-id: 9f6b6308454760669e28550d2edf3e4c42e65ff7

configs/inpainting.yaml

@@ -5,7 +5,7 @@ target_width : 2560
 target_height : 1472
 prompt : 'Product on the table 4k ultrarealistic'
 negative_prompt : 'low resolution , bad resolution , Deformation , Weird Artifacts, bad quality,blown up image, high brightness , high saturation '
-roi_scale : 0.7
+roi_scale : 0.5
 strength : 0.6
 guidance_scale : 7
 num_inference_steps : 150

product_diffusion_api/endpoints.py

@@ -34,14 +34,15 @@ logfire.instrument_fastapi(app)
 @app.get('/')
 async def root():
     return {
-        'message': 'Welcome to the main API Page for the product Photography Application , we provide APIs for building Stunning Product Images with Diffusion Models',
+        'message': 'Welcome to the PICPILOT API Page , Develop Visual Stories for your Brand',
         'description': 'This API provides endpoints for accessing and managing product diffusion data.',
         'version': '1.0.0',
         'author': 'Vikramjeet Singh',
         'contact': {
             'email': 'singh.vikram.1782000@gmail.com',
             'github': 'https://github.com/vikramxD',
-            'website': 'https://vikramxd.github.io'
+            'website': 'https://vikramxd.github.io',
+            'peerlist': 'https://peerlist.io/vikramxd'
         },
         'license': 'MIT',
     }

product_diffusion_api/routers/painting.py

@@ -11,6 +11,8 @@ from PIL import Image
 import io
 from utils import ImageAugmentation
 from hydra import compose, initialize
+import lightning.pytorch as pl
+pl.seed_everything(42)
 
 router = APIRouter()
 
@@ -106,13 +108,26 @@ def run_inference(cfg: dict, image_path: str, prompt: str, negative_prompt: str,
         HTTPException: If an error occurs during the inference process.
 
     """
-    image, mask_image = augment_image(image_path, cfg['target_width'], cfg['target_height'], cfg['roi_scale'], cfg['segmentation_model'], cfg['detection_model'])
+    image, mask_image = augment_image(image_path, 
+                                      cfg['target_width'], 
+                                      cfg['target_height'], 
+                                      cfg['roi_scale'], 
+                                      cfg['segmentation_model'], 
+                                      cfg['detection_model'])
     
-    pipeline = AutoPaintingPipeline(model_name=cfg['model'], image=image, mask_image=mask_image, target_height=cfg['target_height'], target_width=cfg['target_width'])
-    output = pipeline.run_inference(prompt=prompt, negative_prompt=negative_prompt, num_inference_steps=num_inference_steps, strength=strength, guidance_scale=guidance_scale)
+    pipeline = AutoPaintingPipeline(model_name=cfg['model'], 
+                                    image=image, 
+                                    mask_image=mask_image, 
+                                    target_height=cfg['target_height'], 
+                                    target_width=cfg['target_width'])
+    output = pipeline.run_inference(prompt=prompt, 
+                                    negative_prompt=negative_prompt, 
+                                    num_inference_steps=num_inference_steps, 
+                                    strength=strength, 
+                                    guidance_scale=guidance_scale)
     return pil_to_s3_json(output, file_name="output.png")
 
-@router.post("/inpainting")
+@router.post("kandinskyv2.2_inpainting")
 async def inpainting_inference(image: UploadFile = File(...), 
                                prompt: str = "", 
                                negative_prompt: str = "", 
@@ -120,7 +135,7 @@ async def inpainting_inference(image: UploadFile = File(...),
                                strength: float = 0.5, 
                                guidance_scale: float = 7.5):
     """
-    Run the inpainting inference pipeline.
+    Run the inpainting/outpainting inference pipeline.
     """
     try:
         image_bytes = await image.read()

product_diffusion_api/yolov8l.pt.REMOVED.git-id

@@ -0,0 +1 @@
+4b11a02a7599e520f9f14a4703f4991237d6ce50

scripts/inpainting_pipeline.py

@@ -1,15 +1,14 @@
 import torch
-from diffusers import AutoPipelineForInpainting
+from diffusers import AutoPipelineForInpainting,DiffusionPipeline
 from diffusers.utils import load_image
 from utils import (accelerator, ImageAugmentation, clear_memory)
 import hydra
 from omegaconf import DictConfig
 from PIL import Image
-import lightning.pytorch as pl
-#pl.seed_everything(1234)
-
+from functools import lru_cache
 
 
+@lru_cache(maxsize=1)
 class AutoPaintingPipeline:
     """
     AutoPaintingPipeline class represents a pipeline for auto painting using an inpainting model from diffusers.
@@ -41,7 +40,7 @@ class AutoPaintingPipeline:
         Returns:
             Image: The output image after inpainting.
         """
-        clear_memory()
+       
         image = load_image(self.image)
         mask_image = load_image(self.mask_image)
         output = self.pipeline(prompt=prompt,negative_prompt=negative_prompt,image=image,mask_image=mask_image,num_inference_steps=num_inference_steps,strength=strength,guidance_scale=guidance_scale, height = self.target_height ,width = self.target_width).images[0]
@@ -63,7 +62,6 @@ def inference(cfg: DictConfig):
     extended_image = augmenter.extend_image(image)
     mask_image = augmenter.generate_mask_from_bbox(extended_image, cfg.segmentation_model, cfg.detection_model)
     mask_image = augmenter.invert_mask(mask_image)
-    #mask_image = cv2.GaussianBlur(np.array(mask_image), (7,7), 0)
     prompt = cfg.prompt
     negative_prompt = cfg.negative_prompt
     num_inference_steps = cfg.num_inference_steps