anuashok/ocr-captcha-v3

This model is a fine-tuned version of microsoft/trocr-base-printed on Captchas of the type shown below

Training Summary

• CER (Character Error Rate): 0.01394585726004922
• Hyperparameters:
  • Learning Rate: 1.5078922700531405e-05
  • Batch Size: 16
  • Num Epochs: 7
  • Warmup Ratio: 0.14813004670666596
  • Weight Decay: 0.017176551931326833
  • Num Beams: 2
  • Length Penalty: 1.3612823161368288

Usage

from transformers import VisionEncoderDecoderModel, TrOCRProcessor
import torch
from PIL import Image

# Load model and processor
processor = TrOCRProcessor.from_pretrained("anuashok/ocr-captcha-v3")
model = VisionEncoderDecoderModel.from_pretrained("anuashok/ocr-captcha-v3")

# Load image
image = Image.open('path_to_your_image.jpg').convert("RGB")
# Load and preprocess image for display
image = Image.open(image_path).convert("RGBA")
# Create white background
background = Image.new("RGBA", image.size, (255, 255, 255))
combined = Image.alpha_composite(background, image).convert("RGB")

# Prepare image
pixel_values = processor(combined, return_tensors="pt").pixel_values

# Generate text
generated_ids = model.generate(pixel_values)
generated_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
print(generated_text)