refactor: Update import statement for accelerator in sdxl_text_to_image.py

product_diffusion_api/routers/sdxl_text_to_image.py

@@ -13,7 +13,7 @@ from functools import lru_cache
 from s3_manager import S3ManagerService
 from PIL import Image
 import io
-from utils import accelerator
+from scripts.utils import accelerator
 
 device = accelerator()
 torch._inductor.config.conv_1x1_as_mm = True

product_diffusion_api/utils.py

@@ -1,12 +0,0 @@
-import torch
-
-
-
-def accelerator():
-    if torch.cuda.is_available():
-        device = 'cuda'
-    elif torch.backends.mps.is_available():
-        device = 'mps'
-    else :
-        device = 'cpu'
-    return device 

requirements.txt

@@ -19,4 +19,5 @@ tensorboard
 Jinja2
 datasets
 peft
-async-batcher
+async-batcher
+ultralytics

run.sh

@@ -1 +1,2 @@
 apt-get update && apt-get install python3-dev
+pip install -r requirements.txt

scripts/utils.py

@@ -0,0 +1,105 @@
+import torch
+from ultralytics import YOLO
+from transformers import pipeline 
+import cv2
+import numpy as np
+
+
+def accelerator():
+    """
+    Determines the device accelerator to use based on availability.
+
+    Returns:
+        str: The name of the device accelerator ('cuda', 'mps', or 'cpu').
+    """
+    if torch.cuda.is_available():
+        device = 'cuda'
+    elif torch.backends.mps.is_available():
+        device = 'mps'
+    else:
+        device = 'cpu'
+    return device
+
+
+
+
+def center_scaled_roi(image_path, bg_size, scale_factor):
+    """
+    Center and scale the region of interest (ROI) within a background image.
+
+    Args:
+        image_path (str): The path to the original image.
+        bg_size (tuple): The size (width, height) of the background image.
+        scale_factor (float): The scaling factor to apply to the ROI.
+
+    Returns:
+        numpy.ndarray: The background image with the scaled ROI centered.
+
+    """
+
+    original_image = cv2.imread(image_path)
+    height, width = original_image.shape[:2]
+
+    # Convert the image to grayscale
+    gray = cv2.cvtColor(original_image, cv2.COLOR_BGR2GRAY)
+
+    # Apply Gaussian blur to reduce noise
+    blurred = cv2.GaussianBlur(gray, (5, 5), 0)
+
+    # Perform edge detection using Canny
+    edges = cv2.Canny(blurred, 50, 150)
+
+    # Find contours in the edged image
+    contours, _ = cv2.findContours(edges.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
+
+    # Initialize variables to store ROI coordinates
+    roi_x, roi_y, roi_w, roi_h = 0, 0, 0, 0
+
+    # Loop over the contours
+    for contour in contours:
+        # Approximate the contour
+        peri = cv2.arcLength(contour, True)
+        approx = cv2.approxPolyDP(contour, 0.02 * peri, True)
+
+        # If the contour has 4 vertices, it's likely a rectangle
+        if len(approx) == 4:
+            # Get the bounding box of the contour
+            x, y, w, h = cv2.boundingRect(approx)
+            roi_x, roi_y, roi_w, roi_h = x, y, w, h
+            break
+
+    # Calculate dimensions for the background
+    bg_width, bg_height = bg_size
+
+    # Resize the ROI based on the scale factor
+    scaled_roi_w = int(roi_w * scale_factor)
+    scaled_roi_h = int(roi_h * scale_factor)
+
+    # Calculate offsets to center the scaled ROI within the background
+    x_offset = (bg_width - scaled_roi_w) // 2
+    y_offset = (bg_height - scaled_roi_h) // 2
+
+    # Resize the original image
+    scaled_image = cv2.resize(original_image, (scaled_roi_w, scaled_roi_h))
+
+    # Create a blank background
+    background = np.zeros((bg_height, bg_width, 3), dtype=np.uint8)
+
+    # Place the scaled ROI onto the background
+    background[y_offset:y_offset+scaled_roi_h, x_offset:x_offset+scaled_roi_w] = scaled_image
+
+    return background
+
+# Define dimensions for the background (larger than the ROI)
+bg_width, bg_height = 800, 600
+
+# Define the scale factor
+scale_factor = 0.5  # Adjust this value as needed
+
+# Call the function to center the scaled ROI within the background
+centered_scaled_roi = center_scaled_roi('image.jpg', (bg_width, bg_height), scale_factor)
+
+# Display the centered scaled ROI
+cv2.imshow('Centered Scaled ROI', centered_scaled_roi)
+cv2.waitKey(0)
+cv2.destroyAllWindows()