Skip to content

Issues: microsoft/DeepSpeed

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

570 Open 735 Closed
570 Open 735 Closed
Author
Filter by author
Label
Filter by label
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Milestones
Filter by milestone
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

[REQUEST] sync FusedAdam with the upstream enhancement New feature or request
#3006 opened Mar 13, 2023 by stas00
[BUG]Assertion error while using pipeline parallelism bug Something isn't working training
#3003 opened Mar 13, 2023 by x54-729
@GuanhuaWang
[BUG] Peft Training with Zero.Init() and Zero3 will increase GPU memory every forward step bug Something isn't working training
#3002 opened Mar 13, 2023 by dumpmemory
1
[REQUEST] Support Deepspeed inference for Fairseq Transformer LM model enhancement New feature or request inference
#3001 opened Mar 13, 2023 by krishnanNuance
[BUG] NVMe Offloading OOMs While Zero Stage 3 Offloading Runs bug Something isn't working training
#3000 opened Mar 13, 2023 by stanleyshly
1
[BUG] Zero Offload Is Significantly Slower Than Normal Training bug Something isn't working training
#2998 opened Mar 13, 2023 by stanleyshly
1
[BUG] deepspeed_stage_3 was used in pytorch_lightning。when initialize, it cost huge cpu memory which increase with the grow of gpu num。 bug Something isn't working training
#2997 opened Mar 12, 2023 by linyubupa
@ShijieZZZZ
2
13B model training OOM with 8x48G machine and limited CPU RAM bug Something isn't working training
#2996 opened Mar 11, 2023 by lavaaa7
6
[BUG] Does bf16 support Zero stage 1 with pipeline? bug Something isn't working training
#2994 opened Mar 10, 2023 by lyj201002
@jomayeri
1
should DeepSpeedEngine.save_checkpoint be only under main_process bug Something isn't working training
#2993 opened Mar 10, 2023 by better629
1
INT8 model inference issue inference
#2991 opened Mar 10, 2023 by remarkableJ
@lekurile
[BUG] ZeRO-Offload GPU -> CPU datatype bug Something isn't working training
#2986 opened Mar 9, 2023 by taehyunzzz
1
[BUG] DeepSpeed Inference reports Signal code: Integer divide-by-zero when Seq length is 4096 for GPT2 bug Something isn't working inference
#2985 opened Mar 9, 2023 by zhen-jia
@cmikeh2
2
[BUG] Sudden increase in CPU memory usage when engine.step() using zero-3 bug Something isn't working training
#2984 opened Mar 9, 2023 by 4AKker
@samadejacobs
Change request regarding the use of CUDA_VISIBLE_DEVICES in deepspeed/launcher/runner.py enhancement New feature or request
#2980 opened Mar 9, 2023 by JY-Ren
AttributeError: 'PipelineEngine' object has no attribute 'flatten' bug Something isn't working training
#2978 opened Mar 9, 2023 by suiyan538
@tohtana
3
【BUG】trainig use CPU offload raise OOM error in wsl2 system training
#2977 opened Mar 9, 2023 by lpty
2
[BUG] export CUDA_VISIBLE_DEVICES=0,1,6,7 does not work bug Something isn't working training
#2976 opened Mar 9, 2023 by xu-song
@loadams
3
[REQUEST] Custom partition option for PipelineModule enhancement New feature or request
#2974 opened Mar 9, 2023 by sgunasekar
Multi-node training reports "stop_waiting response required" and "connection reset by peer" bug Something isn't working training
#2973 opened Mar 8, 2023 by maxmaier59
[BUG] 'StableDiffusionPipeline' object has no attribute 'children' bug Something isn't working inference
#2968 opened Mar 8, 2023 by stevensu1977
@molly-smith
7
Does deepspeed automaticlly partition the model to multi gpus when stage=3 is specified?
#2966 opened Mar 8, 2023 by superzhangmch
I think this should be isinstance (batch[0], (tuple, list))
#2959 opened Mar 7, 2023 by CoinCheung
[REQUEST]How to deploy multi-nodes training without hostfile? enhancement New feature or request
#2958 opened Mar 7, 2023 by SefaZeng
2
deepspeed on T4 GPU server and run Stable diffustion model inference error bug Something isn't working inference
#2957 opened Mar 7, 2023 by qingyuan18
@lokoppakmsft
1
ProTip! no:milestone will show everything without a milestone.