LoreonLabsPlatform

Overview

Intelligence

Markets
Builders
Research
Narratives
Ecosystems
Launchpads

Discover

Search
Sources

Python

yaml-bert

Transformer with tree positional encoding for learning patterns in structured data (YAMLs)

PythonEmergingbertbigramkubernetespytorch

Stars

1

Forks

—

Contributors

1

Last push

15d ago

Recent commits

Latest commits.

docs(v11): brainstorm — cross-document attention as the next direction
b3e826eVimal Kumar18d ago
docs(v10): prep Task 10 — results doc skeleton + spec pos_weight addendum
58a25bcVimal Kumar18d ago
fix(v10): pos_weight=2400 in recon BCE, recon_weight=0.001 default
27a2b1fVimal Kumar18d ago
docs(v10): smoke training run log — 5K x 3 epochs, GTX 1650, 366s
c942999Vimal Kumar18d ago
feat(v10): aggregator absorbs VALUE-child hiddens (mean pool)
47669b2Vimal Kumar18d ago

fix(v10): test_dataset_subtree assertion uses subword_vocab_size

a43f8ceVimal Kumar19d ago

fix(v10): expand subtree descs with VALUE children for recon target

707d684Vimal Kumar19d ago

feat(v10): dataset emits bag-of-VALUE-subwords recon targets

5f6853aVimal Kumar19d ago

Top contributors

Builders behind this project.