megatron.training#
Description
Pretrain utilities.
Functions
|
|
|
|
|
Evaluation. |
|
Helper function to evaluate and dump results on screen. |
|
Build the model. |
|
Main training program. |
|
Note that this call will sync across all ranks. |
|
|
|
Single training step. |
|
Log training information such as losses, timing, .... |