Fig. 6: Scaling law on r2SCAN data.
From: Cross-functional transferability in foundation machine learning interatomic potentials
a Energy MAE and b Force MAE on the MP-r2SCAN validation set using either Method 4, TL with r2SCAN AtomRef (Transfer, blue) or Method 1, training from scratch (Scratch, orange) methods. Zero training points in Transfer refers to the performance of the GGA/GGA + U pre-trained CHGNet with r2SCAN AtomRef. Linear fits are applied for x > 1000 to demonstrate the neural scaling law, and the coefficients of determination (R2) are shown in the figures.
