Tensor.sparse_categorical_crossentropy()

Returns the sparse categorical crossentropy loss between two tensors.

Usage

from tinygrad.tensor import Tensor
 
a1 = Tensor([1, 2, 3])
a2 = Tensor([1, 2, 3])
a1 = a1.sparse_categorical_crossentropy(a2)
 
print(a1.numpy())

0.6050707

;