Skip to content

Pull requests: apple/axlearn

Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Assignee
Filter by who’s assigned
Sort

Pull requests list

Fix v6e boolean flags
#889 by Ethanlm was merged Dec 13, 2024 Loading…
Add default compiler options for v6e
#887 by samos123 was merged Dec 12, 2024 Loading…
Support v6e (Trillium)
#879 by Ethanlm was merged Dec 11, 2024 Loading…
[bug fix] fix issues with tool use metrics
#878 by fnan was merged Dec 10, 2024 Loading…
Makes Bastion write scheduler history files
#877 by ruomingp was merged Dec 10, 2024 Loading…
pin scikit-learn
#875 by fnan was merged Dec 9, 2024 Loading…
Refactor attention bias/mask/segments.
#873 by apghml was merged Dec 9, 2024 Loading…
Simplifies ScheduleResults.job_verdicts
#872 by ruomingp was merged Dec 9, 2024 Loading…
add tool use intent metrics
#871 by fnan was merged Dec 6, 2024 Loading…
add merge_group to action workflows
#870 by madrob was merged Dec 6, 2024 Loading…
Enable GCP Workload Monitoring
#868 by Perseus14 was closed Dec 4, 2024 Loading…
Support orbax state builder
#866 by hanzhi713 was merged Dec 6, 2024 Loading…
Fix feedforward remat point
#864 by hanzhi713 was merged Dec 2, 2024 Loading…
Addresses Github Action warning.
#862 by markblee was merged Nov 30, 2024 Loading…
Generalizes decoder by taking input batch.
#861 by markblee was merged Nov 30, 2024 Loading…
Unify init and prefill for attention layers.
#860 by markblee was merged Nov 27, 2024 Loading…
[Bug Fix] Fix a nit issue for Jax rollback
#856 by kelvin-zou was merged Nov 25, 2024 Loading…
[Bug fix] Jax rollback for TPU
#855 by kelvin-zou was closed Nov 23, 2024 Loading…
Rollback to Jax 0.4.33
#854 by kelvin-zou was merged Nov 22, 2024 Loading…
Implement ConvXDTranspose
#853 by ds-hwang was merged Nov 25, 2024 Loading…
Fix RLHF slowdown in attention multi steps extend_step.
#849 by ds-hwang was merged Nov 19, 2024 Loading…
ProTip! Follow long discussions with comments:>50.