Update app.py

app.py

@@ -1,46 +1,15 @@
 import gradio as gr
-from transformers import VisionEncoderDecoderModel, TrOCRProcessor,  AutoTokenizer
+from transformers import VisionEncoderDecoderModel, TrOCRProcessor,AutoTokenizer,ViTFeatureExtractor
 from PIL import Image
+import torch
+
 
-# Load the model and processor from Hugging Face
 tokenizer = AutoTokenizer.from_pretrained("paudelanil/trocr-devanagari")
 model = VisionEncoderDecoderModel.from_pretrained("paudelanil/trocr-devanagari")
 feature_extractor = ViTFeatureExtractor.from_pretrained('google/vit-base-patch16-224-in21k')
-
-processor = TrOCRProcessor(feature_extractor=feature_extractor, tokenizer=tokenizer)
-
 device = 'cuda' if torch.cuda.is_available() else 'cpu'
-model.to(device)
-def preprocess_image(image):
-        # Resize while maintaining aspect ratio
-        target_size = (224, 224)
-        original_size = image.size
-
-        # Calculate the new size while maintaining aspect ratio
-        aspect_ratio = original_size[0] / original_size[1]
-        if aspect_ratio > 1:  # Width is greater than height
-            new_width = target_size[0]
-            new_height = int(target_size[0] / aspect_ratio)
-        else:  # Height is greater than width
-            new_height = target_size[1]
-            new_width = int(target_size[1] * aspect_ratio)
-
-        # Resize the image
-        resized_img = image.resize((new_width, new_height))
-
-        # Calculate padding values
-        padding_width = target_size[0] - new_width
-        padding_height = target_size[1] - new_height
-
-        # Apply padding to center the resized image
-        pad_left = padding_width // 2
-        pad_top = padding_height // 2
-        pad_image = Image.new('RGB', target_size, (255, 255, 255))  # White background
-        pad_image.paste(resized_img, (pad_left, pad_top))
-
-        return pad_image
-    
 
+model.to(device)
 def predict(image):
     # Preprocess the image
     image = Image.open(image).convert("RGB")