Skip to content

Issues: stanford-crfm/levanter

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

64 Open 203 Closed
64 Open 203 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

Can not instantiate an infinite EpochDataset from current config access bug Something isn't working
#884 opened Feb 10, 2025 by kothasuhas
3
Current batch indexing does not interact nicely with varying mixtures
#876 opened Feb 3, 2025 by kothasuhas
@kothasuhas
3
jax-metal support appears broken
#864 opened Jan 23, 2025 by jder
1
LM Eval harness scores are bugged in large runs maybe? p1
#860 opened Jan 21, 2025 by dlwh
@dlwh
Evals or other hooks can be skipped if preempted
#855 opened Jan 14, 2025 by dlwh
Run on existing tpu vms (on demand - not queued)
#854 opened Jan 10, 2025 by kathir-ks
1
make new DataLoader work with flexible batch sizes
#839 opened Dec 12, 2024 by dlwh
Add AQT/int8 training
#838 opened Dec 12, 2024 by dlwh
Investigate MaxText's ICI Transpose sharding
#837 opened Dec 12, 2024 by dlwh
Model checkpoints should store their model configs
#781 opened Oct 30, 2024 by dlwh
Initializing of models from checkpoints/pretrained models has gotten a bit crazy
#780 opened Oct 30, 2024 by dlwh
Using docker workflow and --foreground ignores env in .config
#770 opened Oct 19, 2024 by ahmeda14960
Minor fix for docker TPU workflow
#749 opened Sep 26, 2024 by ahmeda14960
1
don't require output_exemplar for user code
#745 opened Sep 25, 2024 by dlwh
Automatically set up wandb workspace for people
#741 opened Sep 24, 2024 by dlwh
1
New Tokenization Pipeline Fails Silently
#736 opened Sep 20, 2024 by ahmeda14960
1
Tensorboard logging breaks if value is of the type "string"
#735 opened Sep 20, 2024 by mhmaqbool
2
Display summary statistics on stdout
#734 opened Sep 20, 2024 by mhmaqbool
1
Crash on GPU in Roberta branch
#727 opened Sep 12, 2024 by dlwh
Reset optimizer state is not working
#726 opened Sep 12, 2024 by ahmeda14960
add sequence packing for evals
#717 opened Sep 4, 2024 by dlwh
Do whatever work to get autocheckpoint working infrastructure p2
#711 opened Aug 28, 2024 by dlwh
More robust and dynamic multi-slicing TPU training
#690 opened Aug 13, 2024 by Ivan-Zhou
4
Ray Cluster Error at tokenizing documents
#667 opened Jul 22, 2024 by Ivan-Zhou
6
Fix Sophia-H when using Splash Attention
#615 opened Jun 5, 2024 by dlwh
ProTip! Exclude everything labeled bug with -label:bug.