google-research

google-research/vmoe

Stars: 728
Forks: 58
Open beginner issues: 0
Indexed issues: 9
Avg merge: 4h 9m
Merged PRs (30d): 3

Dominant language: Jupyter Notebook
License: Apache-2.0
Last GitHub push: Sep 11, 2026
Latest indexed: Sep 16, 2026
Contributing guide: Contributing guide
Code of conduct: No code of conduct
Beginner labels: No beginner labels indexed

Open All

9 open issues indexed

How do I use the checkpoints from Google Cloud? Open

google-research/vmoe#3 · 3 comments · 1 reaction · 0 assignees · Indexed Sep 16, 2026
Question: have you tried patch-expert assignment level for BPR? Open

google-research/vmoe#49 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
module 'vmoe.checkpoints.partitioned' has no attribute 'Mesh' Open

google-research/vmoe#137 · 0 comments · 5 reactions · 0 assignees · Indexed Sep 16, 2026
Loading Saved Checkpoint After Finetuning Open

google-research/vmoe#148 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
How to use other checkpoints to evaluate the model's accuracy on different datasets, including utilizing gs://vmoe_checkpoints/eee_s32_last2_ilsvrc2012_ft_cifar100 to test datasets other than CIFAR-100? Open

google-research/vmoe#160 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Here are some questions about soft MoE Open

google-research/vmoe#176 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Shape mismatch running `vmoe_b16_imagenet21k_randaug_strong_ft_ilsvrc2012` checkpoint Open

google-research/vmoe#182 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
i came here for the code of SPARSE UPCYCLING: TRAINING MIXTURE-OF-EXPERTS FROM DENSE CHECKPOINTS Open

google-research/vmoe#188 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Replace jax.scipy.stats.norm.cdf with jax.scipy.special.ndtr for improved efficiency Open

google-research/vmoe#200 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026

A short digest of beginner-friendly GitHub issues.