Recognition annotations confusion

[Enverrrr]

Hey I want to train SAR recognition model to read license plate. I prepared a dataset and annotated it after that I trained on 'sar_r31_parallel_decoder_toy_dataset.py' however the model print just first part of license plate here an example

and here annotation file
boşluk_label.txt
also I tried without space among the part of text
label.txt

Unfortunately it's not working. In my opinion I annotated wrong way but I'm not sure what can I do?
P.S. I turned of detection algorithm in mmocr/utils/ocr.py

[gaotongxiao]

I tried without space among the part of text label.txt

I think what you did was right. How did the model perform with this annotation?