21 days ago
-
Report figures and tables (full resolution)
-
Report figures and tables (full resolution)
-
Report figures and tables (full resolution)
-
Repository card: folder layout and links
-
Repository card: folder layout and links
22 days ago
-
p2-seed43-sophia: final weights + checkpoints every 0.1x dataset
-
p2-seed43-muon: final weights + checkpoints every 0.1x dataset
-
p2-seed43-lion: final weights + checkpoints every 0.1x dataset
-
p2-seed43-mars: final weights + checkpoints every 0.1x dataset
-
p2-seed43-adamw: final weights + checkpoints every 0.1x dataset
-
p2-sophia: final weights + checkpoints every 0.1x dataset
-
p2-muon: final weights + checkpoints every 0.1x dataset
-
p2-lion: final weights + checkpoints every 0.1x dataset
-
p2-mars: final weights + checkpoints every 0.1x dataset
-
p2-adamw: final weights + checkpoints every 0.1x dataset
-
p2-seed43-sophia: final weights
-
p2-seed43-muon: final weights
-
p2-seed43-lion: final weights
-
p2-seed43-mars: final weights
-
p2-seed43-adamw: final weights
-
p2-sophia: final weights
-
p2-muon: final weights
-
p2-lion: final weights
-
p2-mars: final weights
-
p2-adamw: final weights
-
p1-seed43-dense: final weights + metrics
-
p1-seed43-moe-e4k1: final weights + metrics
-
p1-seed43-moe-e4k2-active: final weights + metrics
-
p1-seed43-moe-s1e3k1: final weights + metrics
-
p1-seed43-moe-e4k2: final weights + metrics
-
Part 3: aligned token accuracy + extended beam timing
-
Part 3 raw generations + metrics
-
p1-dense: final weights + metrics
-
p1-moe-e4k2-active: final weights + metrics
-
p1-moe-e4k2: final weights + metrics
-
p1-moe-s1e3k1: final weights + metrics
-
p1-moe-e4k1: final weights + metrics
-
initial commit
