Hierarchical incremental learning deciphers molecular arrangements in multi-component materials

# Sample datasets

### data format:

>dataSet_count.npy: A numpy array of size N1, consisting of integers from 1 to N1.

>dataSet_dist.npy: A numpy matrix of shape (N1, M1, M1). The size of the Coulombic matrix is M1 × M1.

>dataSet_name.npy: A numpy matrix of shape (N1, M1). Every row contains the atom set names.

### task A

md1

md2

md3

md4

md5

md6

### task B

md1

md2

md3

### task C

md1

md2

md3

### task D

md1

md2

### task E

md1

md2

md3

md4

### task F

md1

md2

md3

### task G

md1

md2