Hello and thank you for your tutorial. I want to use transformer model on my own dataset, but so far I haven't had a good result. My cer is 0.7 and I used Microsoft small printed model for fine tuning. I do not know what the problem is. My training data contains images that have sentence rather than words; do you think using words rather than sentence would help me improve my result?
Hello and thank you for your tutorial. I want to use transformer model on my own dataset, but so far I haven't had a good result. My cer is 0.7 and I used Microsoft small printed model for fine tuning. I do not know what the problem is. My training data contains images that have sentence rather than words; do you think using words rather than sentence would help me improve my result?