lukas-blecher / LaTeX-OCR

pix2tex: Using a ViT to convert images of equations into LaTeX code.
https://lukas-blecher.github.io/LaTeX-OCR/
MIT License
12.7k stars 1.03k forks source link

training LaTeX-OCR from scratch #385

Open liuhongen1234567 opened 4 months ago

liuhongen1234567 commented 4 months ago

Hello, I want to train a model from scratch, but the training according to the configuration file in the code does not reach the accuracy of the official model. May I ask if the encoder in the model, i.e. hybrid vit, uses a pre-trained model on ImageNet? If so, how can I download it?