Gahow Wang 3c88fb7178 fixup! test: T21 — DDP-dropout regression (live under DDP + p=0 bit-identical)

2026-06-18 21:15:05 +08:00

..

xtrain-autodiff

gqa: real grouped-query attention (repeat_kv op + both SDPA paths + wiring + tests)

2026-06-18 01:37:37 +08:00

xtrain-cuda

gqa: real grouped-query attention (repeat_kv op + both SDPA paths + wiring + tests)

2026-06-18 01:37:37 +08:00

xtrain-distributed

fixup! test: T21 — DDP-dropout regression (live under DDP + p=0 bit-identical)

2026-06-18 21:15:05 +08:00

xtrain-model

gqa: fix kv-proj shape test param indices (embed,attn_norm precede wq)

2026-06-18 01:38:42 +08:00

xtrain-optim

perf: make xtrain-cuda a regular dep of xtrain-optim (GPU AdamW)

2026-06-15 16:53:52 +08:00

xtrain-tensor

gqa: real grouped-query attention (repeat_kv op + both SDPA paths + wiring + tests)

2026-06-18 01:37:37 +08:00

xtrain-train

gqa: real grouped-query attention (repeat_kv op + both SDPA paths + wiring + tests)

2026-06-18 01:37:37 +08:00