Azazel0203
commented
2 months ago Hello! Please help me start training
All the actions I've taken:
- Created a new project (ocr_service) in pycharm with a virtual environment
- Cloned the project (git clone https://github.com/Azazel0203/ocr_captcha.git)
- cd ocr_captcha
- pip install -r requirements.txt
- Added six new characters "+", "-", "*", "!", "@", "." to the file "artifact/unique_char.csv", "artifact/char_to_num.json", "artifact/num_to_char.json"
- 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": []}
- Created a train folder in the {project name}/ocr captcha directory
- Added the file "labels.csv" to the train folder
- Created an images folder in the train directory and placed 93 pictures in it
- In the file "data_ingestion.py" I corrected the paths and the cycle
- 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"
- In the file "model_trainer.py" I changed the line with the sizes of the pictures
- I run the file "data_ingestion.py", there are no errors
- 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
- In the file "model_trainer.py" line 15 is highlighted as an error
- 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
You shouldn't manually change the json files in the artifacts. The data-ingestion part does it for you. Also you shouldn't be changing the paths. Only path you should update is the path to the main data (training and testing data). Rest all the things are supposed to run as it is.
Also delete the artifact folder and then create a new folder name artifact to proceed. As your data is of different kind, You would need to create different artifacts.
Roboxkin
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"
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
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