Skip to content

Pull requests: EleutherAI/gpt-neox

New pull request New
24 Open 786 Closed
24 Open 786 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

Update text_generation_utils.py to work with pipe_parallel_size of 0
#1316 opened Oct 30, 2024 by markNZed Loading…
CUDA/Pytorch multiprocessing workaround and test fixes
#1304 opened Oct 14, 2024 by AI-WAIFU • Draft
pytest-forked alternative to get around CUDA/pytorch multiprocessing limitation
#1303 opened Oct 14, 2024 by AI-WAIFU • Draft
Add support for context parallelism
#1299 opened Oct 1, 2024 by bclyang Loading…
1
TE integration via full TransformerLayer
#1297 opened Sep 30, 2024 by tf-nv Loading…
Reinforce PR
#1288 opened Sep 19, 2024 by dmahan93 Loading…
@Quentin-Anthony
1
TransformerEngine Integration
#1282 opened Sep 16, 2024 by aurelion-source Loading…
@aurelion-source
7
Add initial ring flash attention support
#1266 opened Sep 4, 2024 by dmahan93 • Draft
1
Add tensor parallelism for RWKV
#1237 opened Jun 19, 2024 by jahatef Loading…
9
Add lora support
#1225 opened May 20, 2024 by mkerin • Draft
Add Transformer Engine
#1213 opened May 10, 2024 by Quentin-Anthony • Draft
@Quentin-Anthony @aurelion-source
1
Dmoe integration merge-queue This PR is next on the queue to merge
#1210 opened May 6, 2024 by DayOfThePenguin Loading…
3
Adding replay into GPT-NeoX
#1200 opened Apr 13, 2024 by AIproj Loading…
@AIproj
3
Added infinite lr schedules
#1194 opened Mar 25, 2024 by kshitijkg Loading…
Add DS inference
#1130 opened Jan 25, 2024 by yang • Draft
[muP] Rework
#1087 opened Dec 1, 2023 by lintangsutawika Loading…
@Quentin-Anthony
27
Adding AxoNN's 3D tensor parallelism [WIP] feature request New feature or request
#1086 opened Nov 28, 2023 by siddharth9820 • Draft
1 of 3 tasks
@Quentin-Anthony
5
Implement neox_args processing when OMPI_COMM_WORLD_SIZE>1
#1073 opened Nov 7, 2023 by kyuheejang Loading…
1
Fixing MuP
#1061 opened Oct 19, 2023 by marcobellagente93 Loading…
32
Deepspeed benchmarking
#878 opened Apr 11, 2023 by cr458 • Draft
1
Improve our mup implementation
#837 opened Mar 15, 2023 by Quentin-Anthony • Draft
2 tasks
@Quentin-Anthony
1
Adds a script to convert NeoX 2.0 checkpoints to DeepSpeed's universal checkpoint format merge-queue This PR is next on the queue to merge
#836 opened Mar 14, 2023 by dashstander Loading…
@Quentin-Anthony
18
Torchtyping
#600 opened Mar 23, 2022 by slash-under • Draft
1
Add generating batch size
#592 opened Mar 18, 2022 by zhuzilin Loading…
1
8
ProTip! Find all pull requests that aren't related to any open issues with -linked:issue.