search

vsymbol / CUTIE

CUTIE (TensorFlow implementation of Convolutional Universal Text Information Extractor)
155 stars 78 forks source link

Unable to run the main_evaluate_json file #10

Open PriyalNarang opened 4 years ago

PriyalNarang commented 4 years ago

I am unable to evaluate the model with main_evaluate_json file. I keep getting the following error whenever I run the file. PS: Have given the correct arguments for checkpoint path and file, but I unsure of how to modify the save-prefix argument. image

muralimariyappan commented 4 years ago

Assuming you are using e_ckpt_path=../graph/CUTIE/graph change the --save_prefix parameter to ICDAR_json/meals. So your command should look like "python main_evaluate_json.py --save-prefix=ICDAR_json/meals."

Alternatively create the directory as mentioned in your error log and place the checkpoint file there.

muralimariyappan commented 4 years ago

In order to get the checkpoint file, you have to generate input files in the mentioned format from any dataset, I used ICDAR 2019 SROIE dataset which was configured with this code. Then you have to train using the main_train_json.py file. After a week or so you will get the checkpoint file which will be good enough for prediction.

varalakshmi5626 commented 3 years ago

In order to get the checkpoint file, you have to generate input files in the mentioned format from any dataset, I used ICDAR 2019 SROIE dataset which was configured with this code. Then you have to train using the main_train_json.py file. After a week or so you will get the checkpoint file which will be good enough for prediction.

Hi, I have generate the input files using SROIE data set , but i didn't get checkpoint file ,I have given my own path to ckpt_path.

Thank you in advance!

gibotsgithub commented 2 years ago

@muralimariyappan i have created the json files in the required format. i have 400 invoices data. the main_train_json.py gets killed because it utilises all the RAM. did u face this issue? I have 16 gb of ram.

SergheiDinu commented 2 years ago

Basically it depends what u also git on this issue... If you got that you need to reshape that, maybe u just have to change that embedding_size at line 41 main_evaluate_json.py also.. are u sure u have right model to evaluate?