jgrusewski 388f54dd06 feat(dqn): pre-upload training data to GPU, use cached targets in hot loop ...

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-02-28 02:49:10 +01:00

..

dqn

feat(dqn): pre-upload training data to GPU, use cached targets in hot loop

2026-02-28 02:49:10 +01:00

tft

refactor: restructure repo — crates/, bin/, testing/ layout

2026-02-25 11:56:00 +01:00

liquid.rs

refactor: restructure repo — crates/, bin/, testing/ layout

2026-02-25 11:56:00 +01:00

mamba2.rs

refactor: restructure repo — crates/, bin/, testing/ layout

2026-02-25 11:56:00 +01:00

mod.rs

refactor: restructure repo — crates/, bin/, testing/ layout

2026-02-25 11:56:00 +01:00

ppo.rs

feat(ppo): pre-upload rollout states to GPU, eliminate per-step tensor allocs

2026-02-28 02:48:33 +01:00

tft_parquet.rs

refactor(ml): drop dead TFT Parquet loader (-130 lines)

2026-02-27 18:56:04 +01:00

tlob.rs

refactor: restructure repo — crates/, bin/, testing/ layout

2026-02-25 11:56:00 +01:00

validation_metrics.rs

refactor: restructure repo — crates/, bin/, testing/ layout

2026-02-25 11:56:00 +01:00