PG Field Guide
Learn
Techniques
Emerging
PRs
← Back to Architecture
Tied FP16 embeddings
Architecture
Used in
1 PRs
Best BPB
1.1330
Avg BPB
1.1330
Submissions
PR #635
by aryanbhosale
1.1330
Hyperparameters Across PRs
pr_number
parameters
635
—