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.

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

Feat wget ckpt #40

Merged
merged 63 commits into from
Oct 8, 2024
Merged

Feat wget ckpt #40

merged 63 commits into from
Oct 8, 2024

Conversation

samsja
Copy link
Collaborator

@samsja samsja commented Oct 5, 2024

live reco workig

@samsja samsja force-pushed the feat-wget-ckpt branch 2 times, most recently from c71844e to 3b94a44 Compare October 5, 2024 21:55
@samsja samsja marked this pull request as ready for review October 6, 2024 00:11
@samsja samsja requested a review from Jackmin801 October 6, 2024 00:11
src/zeroband/comms.py Outdated Show resolved Hide resolved
src/zeroband/diloco.py Outdated Show resolved Hide resolved
src/zeroband/train.py Outdated Show resolved Hide resolved
src/zeroband/comms.py Outdated Show resolved Hide resolved
src/zeroband/comms.py Show resolved Hide resolved
src/zeroband/comms.py Outdated Show resolved Hide resolved
src/zeroband/comms.py Outdated Show resolved Hide resolved
Comment on lines 223 to 224
if config.ckpt.resume is not None:
# all is inplace
ckpt_manager.load(resume_ckpt_path=config.resume)
ckpt_manager.load(resume_ckpt_path=config.ckpt.resume)

model.train()
if elastic_device_mesh.live_recovery.need_live_recovery:
ckpt_manager.download_and_load_ckpt_from_peers(
elastic_device_mesh.live_recovery.get_adress(world_info.global_rank - 1)
)
elastic_device_mesh.live_recovery.need_live_recovery = False
diloco.fake_step(model)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you need the load from resume_ckpt_path if you are doing live recovery? does it skip all the state live recovered state dict inits and only does dataloader loading?

src/zeroband/train.py Outdated Show resolved Hide resolved
README.md Outdated Show resolved Hide resolved
@samsja
Copy link
Collaborator Author

samsja commented Oct 8, 2024

working fine now Screenshot from 2024-10-08 12-40-21

@samsja samsja requested a review from JohannesHa October 8, 2024 19:52
Copy link
Member

@Jackmin801 Jackmin801 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

NOICE! lgtm

@Jackmin801 Jackmin801 merged commit f5d7158 into main Oct 8, 2024
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants