DualPipe/examples
A-transformer 7cb1e5e632
Return a scalar via mean
Fix loss dimension issue in ref_step function.
Aggregate into a scalar
2025-03-06 21:21:37 +04:00
..
example_dualpipe.py Return a scalar via mean 2025-03-06 21:21:37 +04:00
example_dualpipev.py add dualpipev 2025-03-04 17:50:00 +08:00