Skip to content

Issues: microsoft/Tutel

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

38 Open 26 Closed
38 Open 26 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
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

How expert parameters are distributed in the cluster when using the Tutel framework?
#251 opened Oct 30, 2024 by luuck
1
How to load 32-experts Swin-transformer-moe on a 2-GPU machine.
#248 opened Oct 27, 2024 by ywxsuperstar
8
How to convert checkpoint files that adapt to different distributed world sizes
#246 opened Aug 27, 2024 by swjtulinxi
1
Question regarding the load importance loss calculation
#240 opened Jun 15, 2024 by wangyirui
1
How about the cost of TUTEL features?
#239 opened Jun 6, 2024 by fyang064
1
How to use Megablocks in MoE training
#236 opened Jun 5, 2024 by CSCYQJ
1
[Question] Comparison to FasterMoE
#232 opened Apr 17, 2024 by Guodanding
4
Qs
#230 opened Apr 16, 2024 by zws98
3
How to create a custom expert with tutel?
#226 opened Mar 22, 2024 by zws98
19
[Question] Why use datatype ncclInt8 in nccl_all_to_all_scatter_async.
#220 opened Dec 18, 2023 by cicirori
1
How to implement Fairseq-MoE training checkpoint like Swin-MoE?
#219 opened Nov 10, 2023 by withinmiaov
1
Non-surface function utilities only work for contiguous input data
#218 opened Nov 6, 2023 by lyd126
12
tutel is slower than the naive p2p using 2DH for small scale
#214 opened Aug 27, 2023 by DongyuXu77
3
how to use tutel on Megatron Deepspeed
#207 opened Jul 15, 2023 by wangyuxin87
4
Can this package support the one-gpu machine
#206 opened May 31, 2023 by momo1986
5
Training with Data and Expert Parallelism
#204 opened May 15, 2023 by santurini
11
INTERNAL ASSERT FAILED
#203 opened May 2, 2023 by Qicheng-WANG
5
about compute_location and locations
#201 opened Apr 10, 2023 by adverbial03
1
ImportError: cannot import name 'tutel_custom_kernel' from 'tutel.impls.jit_compiler' environmental issue
#198 opened Mar 30, 2023 by zhaojiancheng007
12
tutel/jit_kernels/sparse.py torch.float16 There is a bug in the calculation: the cuda calculation result is inconsistent with the CPU calculation result and the array is out of bounds invalid This doesn't seem right
#196 opened Mar 8, 2023 by WsqRichards1
1
All2All precision always in fp32
#195 opened Feb 21, 2023 by vchiley
1
How the experts' gradients are handled under data parallelism?
#192 opened Dec 26, 2022 by yzs981130
1
RuntimeError: No such operator tutel_ops::cumsum
#190 opened Nov 14, 2022 by sharkdrop
10
[installation errors] fatal error: nccl.h: No such file or directory
#189 opened Oct 19, 2022 by qianyuzqy
1
Multi-nodes training is much more slower than single node
#187 opened Sep 29, 2022 by YingqingHe
1
ProTip! Add no:assignee to see everything that’s not assigned.