Skip to content

Pull requests: apple/axlearn

New pull request New
Clear current search query, filters, and sorts
40 Open 1,008 Closed
40 Open 1,008 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

Repeat KV heads in Flash Attention
#938 by changlan was merged Jan 21, 2025 Loading…
Explicitly pass module outputs to metrics.
#953 by markblee was merged Jan 28, 2025 Loading…
Add v6e special meshes
#952 by hanzhi713 was merged Jan 27, 2025 Loading…
1
Workaround module outputs being dropped.
#951 by markblee was merged Jan 26, 2025 Loading…
Update LoraFusedQKVLinear
#949 by qdavid1 was merged Jan 27, 2025 Loading…
@markblee
update jax to 0.4.37
#948 by matthew-e-hopkins was merged Jan 27, 2025 Loading…
6
Add link to github issue regarding kubernetes-32.0.0
#947 by Ethanlm was merged Jan 24, 2025 Loading…
Pin kubernetes pip version to 31.0.0 to fix client authentication error
#946 by Ethanlm was merged Jan 24, 2025 Loading…
4
Forward input keys to decoder.
#944 by markblee was merged Jan 23, 2025 Loading…
Legacy flash remat fix
#943 by hanzhi713 was merged Jan 23, 2025 Loading…
Some fixes for flash remat
#942 by hanzhi713 was merged Jan 22, 2025 Loading…
Add GKE A3 Ultra support
#940 by samos123 was closed Apr 4, 2025 Loading…
44
Flash Attention for Neuron
#939 by apoorvtintin was merged Feb 13, 2025 Loading…
57
Allow external positions to be inputed in RoPE embedding layer
#926 by Firenze11 was merged Jan 27, 2025 Loading…
17
AOT compilation for v6e
#937 by changlan was merged Jan 21, 2025 Loading…
Adds mesh rule for a3-megagpu-8g.
#936 by markblee was merged Jan 23, 2025 Loading…
Avoid a top-level import of tokenizers.
#935 by markblee was merged Jan 19, 2025 Loading…
Makes causal lm metrics configurable.
#934 by markblee was merged Jan 21, 2025 Loading…
3
Supports flexible input partition specs.
#933 by markblee was merged Jan 19, 2025 Loading…
5
Enable GCP Workload Monitoring
#932 by Perseus14 was closed Mar 18, 2025 Loading…
3
Add ReadOptions args to _make_autoregressive_inputs
#931 by RsEnts was merged Jan 17, 2025 Loading…
3
AdaptiveLayerNormModulation raises ValueError, instead of assert.
#930 by ds-hwang was merged Jan 16, 2025 Loading…
1
Fix aot compilation with grain inputs.
#929 by markblee was merged Jan 16, 2025 Loading…
Add prefill hidden states as module outputs.
#928 by markblee was merged Jan 16, 2025 Loading…
1
Cache AoT compilation result
#927 by hanzhi713 was merged Jan 16, 2025 Loading…
@ruomingp
15
ProTip! Filter pull requests by the default branch with base:main.