Skip to content

Pull requests: apple/axlearn

New pull request New
47 Open 811 Closed
47 Open 811 Closed
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
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Implements FlashDecoding with Sparsity Support
#899 opened Dec 18, 2024 by hanzhi713 Loading…
1
Special remat for Neuron
#898 opened Dec 17, 2024 by apoorvtintin Loading…
42
use "true" and "false" instead of 0 and 1
#890 opened Dec 12, 2024 by samos123 Loading…
1
[DO-NOT-MERGE] PR encompassing all changes needed to support neuron on Axlearn
#886 opened Dec 11, 2024 by apoorvtintin Loading…
Add meshes and config for TRN2/1 for Fuji models
#885 opened Dec 11, 2024 by apoorvtintin Loading…
7
Input batch sharding strategy BATCH
#884 opened Dec 11, 2024 by apoorvtintin Loading…
2
Flash Attention for Neuron
#883 opened Dec 11, 2024 by apoorvtintin Loading…
7
Docker: Upgrade Jax to 0.4.37
#880 opened Dec 10, 2024 by samos123 • Draft
Optimizer offloading through weight-only offload
#867 opened Dec 4, 2024 by hanzhi713 Loading…
47
improve GCS perf: Change resource limit to request
#851 opened Nov 19, 2024 by samos123 Loading…
1
Add llama 3 tokenizer
#850 opened Nov 19, 2024 by sychen52 Loading…
25
Optimize TPU Flash Attention (400x speed-up on 32k long context)
#845 opened Nov 18, 2024 by ds-hwang • Draft
Add Mamab2 and its Jamba variant
#839 opened Nov 14, 2024 by berlino Loading…
2
Integrate Orbax's emergency checkpoint.
#820 opened Nov 7, 2024 by hanzhi713 • Draft
Add Goodput & Badput recording and monitoring support.
#783 opened Oct 25, 2024 by dipannita08 Loading…
5 tasks done
31
restore test_apply_paddings_check runtime_checks test
#771 opened Oct 22, 2024 by mattjj Loading…
5
add Fuji v3 405b and solve HBM OOMs for larger models
#766 opened Oct 21, 2024 by samos123 • Draft
17
Alternative method of submitting jobs to DF Runner
#756 opened Oct 17, 2024 by remylouisew Loading…
3
Converts checkpointer and state builder to configurable.
#751 opened Oct 14, 2024 by markblee • Draft
Create custom jobset for pathways
#746 opened Oct 9, 2024 by jesus-orozco • Draft
5
Use regex for parsing step_dir
#739 opened Oct 7, 2024 by nlusskin Loading…
4
Add --megascale_abort_on_hangs flag for multi-slice TPU jobs
#731 opened Oct 3, 2024 by mugithi Loading…
12
Set TF_FORCE_GPU_ALLOW_GROWTH=true by default
#712 opened Sep 24, 2024 by samos123 Loading…
2
Fix submission of Dataflow jobs
#711 opened Sep 24, 2024 by damccorm Loading…
2
Adding support for Pathways proxy
#690 opened Sep 9, 2024 by jesus-orozco Loading…
22
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.