Machine learning based predictions of the glass transition temperature $$\left(T_\text{g} \right)$$, melting temperature $$\left(T_\text{m}\right)$$, and degradation temperature $$\left(T_\text{d}\right)$$ for copolymers of two comonomers. Homopolymer predictions are possible by leaving the Comonmer B input field empty. The $$2\sigma$$ uncertainties are computed using Monte Carlo dropouts.