-
Notifications
You must be signed in to change notification settings - Fork 280
Pull requests: apple/axlearn
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Supports configuring causal=True optimization in FlashAttention mha.
#31
by markblee
was merged Aug 7, 2023
Loading…
Supports retaining intermediate output collections across scan.
#56
by markblee
was merged Sep 2, 2023
Loading…
Supports stack-of-stacks and repeat-of-stacks in transformer decoding.
#58
by markblee
was merged Sep 6, 2023
Loading…
Supports configuring instances as dst_layer in hf builder.
#71
by markblee
was merged Sep 14, 2023
Loading…
Introduces SpmdTrainer._train_step_input_partition_specs()
#202
by ruomingp
was merged Nov 25, 2023
Loading…
ProTip!
Exclude everything labeled
bug
with -label:bug.