azazel0203 / ocr_captcha

This project creates a basic web service for solving image-based CAPTCHAs. Using the Flask framework, it allows users to upload CAPTCHA images and employs an Optical Character Recognition (OCR) pipeline to extract the embedded text.

---

---

📖 Education

• Bachelor's in Electronics
• Indian Institute of Technology - Varanasi, Graduate in 2025

👤 About Me

• 🌱 I'm currently exploring the realms of Machine Learning, Generative AI, and Natural Language Processing.
• 🔭 I’m passionate about leveraging AI to solve real-world problems and create innovative solutions.
• 📫 Feel free to reach out to me at

🛠️ Skills

• Programming Languages:
• Competitive Programming: - Expert
• Machine Learning: ML and DL Algorithms (GANS, VAEs, CNNs, Neural Networks)
• Deep Learning: Deep Reinforcement Learning, Artificial Neural Networks
• Libraries & Frameworks:

🖥️ Project Highlights

Machine Learning	PyPi Packages
Title Technologies ocr_captcha Sentiment Analyser Hugging Face Invoice Extractor Medical_ChatBot	Title Technologies mongo_only

📈 Github Stats

📊 Contributions

ocr_captcha's Issues

error message ['function' object has no attribute 'Model']

Hello!
Please help me start training

All the actions I've taken:

1. Created a new project (ocr_service) in pycharm with a virtual environment

2. Cloned the project (git clone https://github.com/Azazel0203/ocr_captcha.git)

3. cd ocr_captcha

4. pip install -r requirements.txt

5. Added six new characters "+", "-", "*", "!", "@", "." to the file "artifact/unique_char.csv", "artifact/char_to_num.json", "artifact/num_to_char.json"

6. In the file "artifact/char_to_num.json" and "artifact/num_to_char.json" corrected "vocabulary_size": 32 to "vocabulary_size": 38

unique_char.csv file:

This is how I have my files now "artifact/char_to_num.json" and "artifact/num_to_char.json"

{"config": {"name": "string_lookup_1", "trainable": true, "dtype": "float32", "invert": true, "max_tokens": null, "num_oov_indices": 1, "oov_token": "[UNK]", "mask_token": null, "output_mode": "int", "sparse": false, "pad_to_max_tokens": false, "idf_weights": null, "vocabulary": ["[UNK]", "2", "3", "4", "5", "6", "7", "8", "9", "A", "B", "C", "D", "E", "F", "G", "H", "J", "K", "M", "N", "P", "Q", "R", "S", "T", "U", "V", "W", "X", "Y", "Z", "+", "-", "*", "!", "@", "."], "vocabulary_size": 38, "encoding": "utf-8"}, "weights": []}

7. Created a train folder in the {project name}/ocr captcha directory

8. Added the file "labels.csv" to the train folder
   

9. Created an images folder in the train directory and placed 93 pictures in it
   

10. In the file "data_ingestion.py" I corrected the paths and the cycle
    

11. Manually added the necessary data to the files (I don't know how correct this is) "raw_img_path.csv", "raw_labels.csv", "test_x.csv", "test_y.csv", "train_x.csv", "train_y.csv"

12. In the file "model_trainer.py" I changed the line with the sizes of the pictures
    

13. I run the file "data_ingestion.py", there are no errors

14. I run the file "model_trainer.py" and get errors

D:\ocr_service\.venv\Scripts\python.exe D:\ocr_service\ocr_captcha\src\ocr_captcha\components\model_trainer.py 
2024-05-07 09:25:38.677735: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2024-05-07 09:25:39.712543: I tensorflow/core/util/port.cc:113] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2024-05-07 09:25:41.673057: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
Traceback (most recent call last):
  File "D:\ocr_service\ocr_captcha\src\ocr_captcha\components\model_trainer.py", line 82, in initate_model_training
    model = build_model(self.img_width, self.img_height, char_to_num) 
  File "D:\ocr_service\ocr_captcha\src\ocr_captcha\utils\utils.py", line 178, in build_model
    model = keras.models.Model(
AttributeError: 'function' object has no attribute 'Model'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "D:\ocr_service\ocr_captcha\src\ocr_captcha\components\model_trainer.py", line 129, in <module>
    model_path = trainer.initate_model_training(train_path_x, train_path_y, test_path_x, test_path_y, unique_chars)
  File "D:\ocr_service\ocr_captcha\src\ocr_captcha\components\model_trainer.py", line 119, in initate_model_training
    raise customexception(e, sys)
ocr_captcha.src.ocr_captcha.exception.exception.customexception: Error occured in python script | 
 name [D:\ocr_service\ocr_captcha\src\ocr_captcha\components\model_trainer.py] 
 line number [82] 
 error message ['function' object has no attribute 'Model']

Process finished with exit code 1

15. In the file "model_trainer.py" line 15 is highlighted as an error
    

16. You have written "inference/: Module for making predictions on new images.", there is no such file

What am I doing wrong?
Please help me start training and then test the model.
I would be grateful for any help!
Thank you

Python version 3.10.11
Windows 11 x64
no GPU

How to correctly generate training data?

Hello!
How to correctly generate training data?

Please tell us in more detail what files are needed and what structure
Thank you

P.S. Not all dependencies are installed
P.P.S. labels? How to create them correctly?