Skip to content
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

Jump to bottom

autotp training(fix dco) #7004

Merged
merged 1 commit into from
Feb 5, 2025
Merged

autotp training(fix dco) #7004

merged 1 commit into from
Feb 5, 2025

Conversation

inkcherry
Copy link
Contributor

@inkcherry inkcherry commented Feb 5, 2025 •
edited
Loading

Same as this PR. affeb88
I noticed the CI updated the DCO check recently. Using the suggested rebase method for sign-off would reintroduce many conflicts, so I opted for a squash merge with sign-off instead. thanks: )

@inkcherry
autotp training squash merge
09d09a5 
Signed-off-by: inkcherry <mingzhi.liu@intel.com>
@tjruwase tjruwase added this pull request to the merge queue Feb 5, 2025
Merged via the queue into deepspeedai:master with commit f04649d Feb 5, 2025
12 checks passed
@delock
Copy link
Collaborator

delock commented Feb 6, 2025

Kudos @inkcherry for contributing AutoTP training! It's a nice feature make tensor parallel training/finetuning more available to HF model users.

I think a tutorial page would help user discover and learn how to use this feature in DeepSpeed. Is it possible to write a tutorial and add it under https://github.com/deepspeedai/DeepSpeed/tree/master/docs/_tutorials introducing steps how to use this feature? I remember you have an example training alpaca with DeepSpeed AutoTP.

tjruwase pushed a commit that referenced this pull request Feb 6, 2025
@inkcherry @tjruwase
autotp training(fix dco) (#7004)
64c4b04 
Same as [this PR](#6922).
[affeb88](affeb88)
I noticed the CI updated the DCO check recently. Using the suggested
rebase method for sign-off would reintroduce many conflicts, so I opted
for a squash merge with sign-off instead. thanks: )

Signed-off-by: inkcherry <mingzhi.liu@intel.com>
Signed-off-by: Olatunji Ruwase <olruwase@microsoft.com>
fitzjalen pushed a commit to fitzjalen/DeepSpeed that referenced this pull request Feb 6, 2025
@inkcherry @fitzjalen
autotp training(fix dco) (deepspeedai#7004)
84beed3 
Same as [this PR](deepspeedai#6922).
[affeb88](deepspeedai@affeb88)
I noticed the CI updated the DCO check recently. Using the suggested
rebase method for sign-off would reintroduce many conflicts, so I opted
for a squash merge with sign-off instead. thanks: )

Signed-off-by: inkcherry <mingzhi.liu@intel.com>
siqi654321 pushed a commit to siqi654321/DeepSpeed that referenced this pull request Feb 7, 2025
@inkcherry
autotp training(fix dco) (deepspeedai#7004)
b6e37a2 
Same as [this PR](deepspeedai#6922).
[affeb88](deepspeedai@affeb88)
I noticed the CI updated the DCO check recently. Using the suggested
rebase method for sign-off would reintroduce many conflicts, so I opted
for a squash merge with sign-off instead. thanks: )

Signed-off-by: inkcherry <mingzhi.liu@intel.com>
Signed-off-by: siqi <siqi@tecorigin.com>
@inkcherry
Copy link
Contributor Author

Kudos @inkcherry for contributing AutoTP training! It's a nice feature make tensor parallel training/finetuning more available to HF model users.

I think a tutorial page would help user discover and learn how to use this feature in DeepSpeed. Is it possible to write a tutorial and add it under https://github.com/deepspeedai/DeepSpeed/tree/master/docs/_tutorials introducing steps how to use this feature? I remember you have an example training alpaca with DeepSpeed AutoTP.

Yes, I will add some document soon~

@tjruwase
Copy link
Contributor

tjruwase commented Feb 7, 2025

@inkcherry, I think a blog would be appropriate to publicize this amazing technology. Although blogs can be a bit work, we will be glad to collaborate and jointly advertise.

@tjruwase tjruwase mentioned this pull request Feb 7, 2025
Closed
loadams pushed a commit that referenced this pull request Feb 7, 2025
@inkcherry @loadams
autotp training(fix dco) (#7004)
d33acaf 
Same as [this PR](#6922).
[affeb88](affeb88)
I noticed the CI updated the DCO check recently. Using the suggested
rebase method for sign-off would reintroduce many conflicts, so I opted
for a squash merge with sign-off instead. thanks: )

Signed-off-by: inkcherry <mingzhi.liu@intel.com>
Signed-off-by: Logan Adams <loadams@microsoft.com>
traincheck-team pushed a commit to traincheck-team/DeepSpeed that referenced this pull request Feb 9, 2025
@inkcherry @traincheck-team
autotp training(fix dco) (deepspeedai#7004)
397b654 
Same as [this PR](deepspeedai#6922).
[affeb88](deepspeedai@affeb88)
I noticed the CI updated the DCO check recently. Using the suggested
rebase method for sign-off would reintroduce many conflicts, so I opted
for a squash merge with sign-off instead. thanks: )

Signed-off-by: inkcherry <mingzhi.liu@intel.com>
gyou2021 pushed a commit to gyou2021/DeepSpeed that referenced this pull request Feb 18, 2025
@inkcherry @gyou2021
autotp training(fix dco) (deepspeedai#7004)
3ea60f0 
Same as [this PR](deepspeedai#6922).
[affeb88](deepspeedai@affeb88)
I noticed the CI updated the DCO check recently. Using the suggested
rebase method for sign-off would reintroduce many conflicts, so I opted
for a squash merge with sign-off instead. thanks: )

Signed-off-by: inkcherry <mingzhi.liu@intel.com>
gyou2021 pushed a commit to gyou2021/DeepSpeed that referenced this pull request Feb 18, 2025
@inkcherry @gyou2021
autotp training(fix dco) (deepspeedai#7004)
f583d30 
Same as [this PR](deepspeedai#6922).
[affeb88](deepspeedai@affeb88)
I noticed the CI updated the DCO check recently. Using the suggested
rebase method for sign-off would reintroduce many conflicts, so I opted
for a squash merge with sign-off instead. thanks: )

Signed-off-by: inkcherry <mingzhi.liu@intel.com>
Signed-off-by: gyou2021 <ganmei.you@intel.com>
@inkcherry inkcherry mentioned this pull request Feb 25, 2025
Closed
@oelayan7
Copy link
Contributor

@inkcherry @loadams @tjruwase
cc @nelyahu
Please notice that this PR has removed keep_module_on_host which was added in #6846

keep_module_on_host is currently ignored and not used in auto_tp.
Please align the current code to the old version that was in auto_tp

@tjruwase
Copy link
Contributor

@inkcherry, can you please help address issues raised by @oelayan7. It seems we need to re-apply the lost changes from #6846.

@inkcherry
Copy link
Contributor Author

inkcherry commented Feb 27, 2025 •
edited
Loading

@oelayan7
Sorry for the break, I actually noticed this issue as well.
You can use this as a temporary workaround for now:
in deepspeed/module_inject/layers set
_partition = move(_partition, 'cpu').detach()

I will fix it later. Thanks!

gyou2021 pushed a commit to gyou2021/DeepSpeed that referenced this pull request Feb 28, 2025
@inkcherry @gyou2021
autotp training(fix dco) (deepspeedai#7004)
18c712f 
Same as [this PR](deepspeedai#6922).
[affeb88](deepspeedai@affeb88)
I noticed the CI updated the DCO check recently. Using the suggested
rebase method for sign-off would reintroduce many conflicts, so I opted
for a squash merge with sign-off instead. thanks: )

Signed-off-by: inkcherry <mingzhi.liu@intel.com>
tohtana pushed a commit that referenced this pull request Feb 28, 2025
@inkcherry @tohtana
autotp training(fix dco) (#7004)
a896693 
Same as [this PR](#6922).
[affeb88](affeb88)
I noticed the CI updated the DCO check recently. Using the suggested
rebase method for sign-off would reintroduce many conflicts, so I opted
for a squash merge with sign-off instead. thanks: )

Signed-off-by: inkcherry <mingzhi.liu@intel.com>
Signed-off-by: Masahiro Tanaka <mtanaka@microsoft.com>
ys950902 pushed a commit to ys950902/DeepSpeed that referenced this pull request Mar 6, 2025
@inkcherry @ys950902
autotp training(fix dco) (deepspeedai#7004)
2101fe3 
Same as [this PR](deepspeedai#6922).
[affeb88](deepspeedai@affeb88)
I noticed the CI updated the DCO check recently. Using the suggested
rebase method for sign-off would reintroduce many conflicts, so I opted
for a squash merge with sign-off instead. thanks: )

Signed-off-by: inkcherry <mingzhi.liu@intel.com>
Signed-off-by: yisheng <yi.sheng@intel.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@tjruwase tjruwase tjruwase approved these changes

@loadams loadams Awaiting requested review from loadams loadams is a code owner

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@inkcherry @delock @tjruwase @oelayan7