Update experiments.py

experiments.py

@@ -4,6 +4,161 @@ import phoenix_helpers
 import helpers 
 
 
+
+import os
+abspath = os.path.abspath('') ## String which contains absolute path to the script file
+os.chdir(abspath) ## Setting up working directory
+import copy
+
+from transformers import AutoProcessor, AutoModelForCausalLM
+from PIL import Image
+import requests
+import utils
+from PIL import Image, ImageDraw, ImageFont
+
+image = Image.open(r"C:\Users\thrin\Downloads\rohitha\floerence.png").convert("RGB")
+
+
+model_id = 'microsoft/Florence-2-large'
+model = AutoModelForCausalLM.from_pretrained(model_id, trust_remote_code=True).eval().cuda()
+processor = AutoProcessor.from_pretrained(model_id, trust_remote_code=True)
+
+utils.set_model_info(model, processor)
+
+image
+
+img_list = ["/content/images_ocr_image_3.jpg",
+ "/content/images_ocr_image_2.jpg",
+ "/content/images_ocr_image_1.jpg",
+ "/content/images_ocr_image_4.png",
+ "/content/pci complliance and Credit card Authorization forms Hand written - Copy.jpg",
+ "/content/Charlemagne.png",
+ "/content/Frederick II.png",
+ "/content/Henry VIII.png",
+ "/content/Louis XIV.png",
+ "/content/William IV.png",
+ "/content/pci complliance and Credit card Authorization forms Hand written - Copy.jpg"
+]
+img_list = [r"C:\Users\thrin\Downloads\rohitha\floerence.png",
+r"C:\Users\thrin\Downloads\rohitha\credit-card-auth-form-xout.jpg"]
+
+for item in img_list:
+    path = item
+    print("<================>")
+    image = Image.open(path)
+    image_rgb = Image.open(path).convert("RGB")
+    
+    tasks = [utils.TaskType.CAPTION,
+           utils.TaskType.DETAILED_CAPTION,
+           utils.TaskType.MORE_DETAILED_CAPTION,]
+    
+    for task in tasks:
+      results = utils.run_example(task, image_rgb)
+      print(f'{task.value}{results[task]}')
+    
+    task = utils.TaskType.OCR
+    results = utils.run_example(task, image_rgb)
+    print('Detected Text: ', results[task])
+
+
+task = utils.TaskType.OCR
+results = utils.run_example(task, image_rgb)
+print('Detected Text: ', results[task])
+
+
+task = utils.TaskType.OCR_WITH_REGION
+results = utils.run_example(task, image_rgb)
+
+# Boxes drawn directly to image, so copy to avoid adulterating image for later tasks
+image_copy = copy.deepcopy(image)
+utils.draw_ocr_bboxes(image_copy, results[task])
+
+
+
+task = utils.TaskType.OCR
+results = utils.run_example(task, image_rgb)
+print('Detected Text: ', results[task])
+
+task = utils.TaskType.OCR_WITH_REGION
+results = utils.run_example(task, image_rgb)
+print('Detected Text: ', results[task])
+
+counter = 0
+for item in img_list:
+    path = item
+    print("<======>",counter,"<======>")
+    image = Image.open(path)
+    image_rgb = Image.open(path).convert("RGB")
+    
+    tasks = [utils.TaskType.CAPTION,
+           utils.TaskType.DETAILED_CAPTION,
+           utils.TaskType.MORE_DETAILED_CAPTION,]
+    
+    for task in tasks:
+      results = utils.run_example(task, image_rgb)
+      print(counter,"<======>",f'{task.value}{results[task]}')
+    
+    task = utils.TaskType.OCR
+    results = utils.run_example(task, image_rgb)
+    print(counter,"<======>",'Detected Text: ', results[task])
+    
+    task = utils.TaskType.OCR_WITH_REGION
+    results = utils.run_example(task, image_rgb)
+    print(counter,"<======>",'Detected Text: ', results[task])
+    
+    
+results[task].items()
+
+results[task].keys()
+
+len(results[task]["quad_boxes"])
+len(results[task]["labels"])
+
+OCR = '<OCR>'
+""" OCR for entire image """
+OCR_WITH_REGION = '<OCR_WITH_REGION>'
+
+from PIL import ImageDraw, Image
+
+{'quad_boxes': [
+    [78.91199493408203,
+     249.8040008544922,
+     332.35198974609375,
+     251.7480010986328,
+     332.35198974609375,
+     299.70001220703125,
+     78.91199493408203,
+     297.7560119628906]
+    ],
+    'labels': ['</s>3702692432']}
+
+def draw_multiple_bounding_boxes(image, coords_and_labels):
+   draw = ImageDraw.Draw(image)
+   width, height = image.size
+   for obj in coords_and_labels:
+       # Extract the bounding box coordinates
+       y1, x1, y2, x2 = obj['bbox'][0] * height, obj['bbox'][1] * width, obj['bbox'][2] * height, obj['bbox'][3] * width
+       # Draw bounding box and label
+       draw.rectangle([x1, y1, x2, y2], outline="red", width=3)
+       draw.text((x1, y1), obj['label'], fill="red")
+   return image
+
+image  = r"C:\Users\thrin\Downloads\rohitha\floerence.png"
+image  = Image.open(image)
+
+quad_box =  [(78.91199493408203,249.8040008544922),
+             (332.35198974609375,251.7480010986328),
+             (332.35198974609375,299.70001220703125),
+             (78.91199493408203,297.7560119628906)]
+
+draw = ImageDraw.Draw(image)
+
+draw.polygon(quad_box,outline="green",width=2)
+image.show()
+image.save(r"C:\Users\thrin\Downloads\florence-2-master (1)\florence-2-master/florence_bounding.png")
+# 
+
+
 load_dotenv()
 models = helpers.fetch_models()