PG Field Guide
Learn
Techniques
Emerging
PRs
← Back to Architecture
mini-MoE
Architecture
Used in
1 PRs
Best BPB
1.1527
Avg BPB
1.1527
Submissions
PR #1228
by meinlebenswerk
1.1527
Hyperparameters Across PRs
pr_number
parameters
1228
{"experts":1}