Request for HME100K Training Settings

[BFlameSwift]

Hi, thank you for the amazing work on PosFormer! I noticed that the model performs very well on both the CROHME and HME100K handwritten mathematical expression datasets. However, I couldn’t find any information on the training settings used for HME100K in the repository or documentation.

Could you please share the training settings or configurations you used for HME100K? This information would be extremely helpful for those of us looking to reproduce and further explore your results on this dataset.

Thanks again for your contributions to HMER research, and I look forward to your response!