triton-lang/triton

how to implement backward propagation

Aperta

#1189 aperta il 13 feb 2023

 (4 commenti) (0 reazioni) (0 assegnatari)MLIR (3136 fork)github user discovery
help wanted

Metriche repository

Star
 (19.995 stelle)
Metriche merge PR
 (Merge medio 2g 18h) (185 PR mergiate in 30 g)

Descrizione

Hi. I am new to triton and cuda. From my understanding, when we implement a customized pytorch operator using cuda, we need to define both forward and backward function, so that the gradients are propagated properly. However, in the triton document (e.g., how to implement vector addition, fused softmax ...), there seems nothing about backward function? Does that mean the backward function is automatically generated by triton? Thx!

Guida contributor