Upload 2 files

app.py

@@ -0,0 +1,56 @@
+import gc
+import torch
+from transformers import TrOCRProcessor, VisionEncoderDecoderModel
+from craft_text_detector import Craft
+from PIL import Image
+import cv2
+import time
+import gradio as gr
+
+# Force CPU usage, disable CUDA
+torch.set_default_device('cpu')
+craft = Craft(output_dir=None, crop_type="box", cuda=False)
+
+# Load smaller model suitable for CPU
+processor = TrOCRProcessor.from_pretrained('microsoft/trocr-small-handwritten')
+model = VisionEncoderDecoderModel.from_pretrained('microsoft/trocr-small-handwritten')
+
+def recognize_handwritten(image):
+    start_time = time.time()
+    
+    # Convert Gradio image to OpenCV format
+    image = cv2.cvtColor(np.array(image), cv2.COLOR_RGB2BGR)
+    result = craft.detect_text(image=image)
+    boxes = result["boxes"]
+    pil_image = Image.fromarray(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
+    texts = []
+    
+    for box in boxes:
+        crop = pil_image.crop([box[0][0], box[0][1], box[2][0], box[2][1]])
+        pixel_values = processor(crop, return_tensors="pt").pixel_values
+        with torch.no_grad():
+            generated_ids = model.generate(pixel_values)
+        text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
+        texts.append(text)
+    
+    text_data = " ".join(texts)
+    end_time = time.time()
+    time_difference = end_time - start_time
+    
+    return f"Recognized text: {text_data}\nTime: {time_difference} seconds"
+
+# Create Gradio interface
+interface = gr.Interface(
+    fn=recognize_handwritten,
+    inputs=gr.Image(type="pil"),
+    outputs="text",
+    title="Handwritten Text Recognition",
+    description="Upload an image containing handwritten text to recognize it."
+)
+
+# Launch the app
+interface.launch()
+
+# Cleanup
+craft.unload_craftnet_model()
+gc.collect()

requirements.txt

@@ -0,0 +1,6 @@
+torch
+transformers
+craft-text-detector
+opencv-python
+Pillow
+gradio