Mahmuod Aboalhassan

I ask if I can get the word or sentence bbox

> is seems that you when load the config file in the first time and start training you got error and then when solve it start the training again if...

thanx it is work

may be when train use the vgg backbone not the craft * update solve is by adding * net.load_state_dict(copyStateDict(torch.load(trained_model, map_location=device)['cuda'])) * in [this line](https://github.com/JaidedAI/EasyOCR/blob/0560fb59b0c280666ded24e533899638ec56e11a/easyocr/detection.py#L85)

I Use this [instructions](https://github.com/PaddlePaddle/PaddleOCR/blob/release/2.6/ppstructure/layout/README.md) the training command is: !export CUDA_VISIBLE_DEVICES=0 !python3 tools/train.py \ -c configs/picodet/legacy_model/application/layout_analysis/picodet_lcnet_x1_0_layout.yml \ --eval and export inference model command is : !python3 tools/export_model.py \ -c configs/picodet/legacy_model/application/layout_analysis/picodet_lcnet_x1_0_layout.yml \...

@LDOUBLEV LDOUBLEV any help

when i run the eval command it runs normally without any error and also the inference command

@LDOUBLEV can you provide a sample of the Arabic dataset you used in the training

if you get any good results than the model of this repo please let me know

[instructions](https://github.com/JaidedAI/EasyOCR/blob/master/custom_model.md) this is the instructions in training a new model and you can fine tune the existing Arabic Model i also train as you and there is no any progress...