LaTeX-OCR
LaTeX-OCR copied to clipboard
Model cannot handle lg or triple root
Original img
model pred:
LaTeX Code: \sqrt[3]{(\\lg 50-1{)}^{3}}- \sqrt{(\lg 2-1{)}^{2}}=
What should I do to improve the model? 😥
Should I fine tune the model with larger dataset?
Really need your help! Thanks! 🙏🏻
I tried feeding more standard images into the model, but instead the model failed to recognise the formulae correctly.😥
model output:
\begin{array}{c c c c c}{{\varepsilon_{\bar{\vee}}\,4\sqrt{\left(\left|\,\vphantom{\cal G}\right.\ \!\lbrack\,\stackrel{\dagger}{\m}\rbrack\,\stackrel{\dagger}{\cal\sqrt}}\longrightarrow}}&{{\sqrt{\left(\left|\,\vphantom{\cal G}\right.\stackrel{\Sigma}{\Sigma}}\stackrel{\Sigma}{\longrightarrow}\stackrel{\Sigma}{\bf\Delta}}}&{{\longrightarrow}}&{{\sqrt{\left(\left|\,\left|\,\vphantom{\cal G}\right.\stackrel{\Sigma}{\cal\Delta}\right.\stackrel{\cal\Delta}\right.\longrightarrow}}}&{{\longrightarrow}}\end{array}