xtrain

Files

Gahow Wang 15f1e526c7 train: parameterize model size (scaling ladder)

Add Config::from_arch(vocab, n_heads, head_dim, n_layers, ffn) so the model
size is a tunable rung instead of a hardcoded tiny config, and Config::core_params()
(num_params minus the two vocab×dim tables) — the figure the ladder is sized
against (the 50257-vocab embed+lm_head adds a fixed ~25M that is not capacity).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

2026-06-15 18:34:39 +08:00

xtrain-autodiff

ops: grad-check the T5 structural ops

2026-06-15 16:05:20 +08:00

xtrain-cuda

perf: streams / drop per-op sync

2026-06-15 16:56:17 +08:00

xtrain-distributed

dist: lengthen scaling bench so NCCL init amortizes

2026-06-15 17:18:23 +08:00

xtrain-model

train: parameterize model size (scaling ladder)