Train the smallest LM you can that fits in 16MB. Best model wins!
Stars
1
Forks
0
Watchers
1
Open Issues
1
Overall repository health assessment
No package.json found
This might not be a Node.js project
41
commits
14
commits
2
commits
1
commits
1
commits
1
commits
1
commits
1
commits
1
commits
1
commits
feat: competitive stack — int6 QAT, BigramHash, SmearGate, OrthoInit
820cb66View on GitHubfeat: verbose training logs — grad norm, lr, tok/s, log every 50 steps
1f943d5View on GitHubfix(rpod): double timer to cover training + eval (~20 min total)
b6c2724View on GitHubfeat(rpod): add preflight check + auto-install gh on pod restart
7d47133View on GitHubfeat: add test suite T6-T30 + fix T16 range for random weights
d250db3View on GitHubfeat(funnel): prototype Funnel Transformer with progressive pooling
b0e572eView on GitHubchore(todos): add latent space language model research item
add01daView on GitHubMerge branch 'exp/ipa-baseline' of https://github.com/nalediym/parameter-golf into exp/ipa-baseline
80428feView on GitHub