jgrusewski 21abd2b0f9 fix: evaluate_baseline bf16→f32 remnants, delete old fxcache files ...

- Fix f32::from_f32/f32::ZERO (nonexistent) in evaluate_baseline.rs
- Eliminate redundant Vec copies (upload source data directly)
- Fix comments referencing __nv_bfloat16 in reductions.rs, training_guard_kernel.cu
- Delete 3 old bf16-format .fxcache files from test_data/

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-04-10 18:57:36 +02:00

..

baseline_common

fix(ml): wire spread_cost_bps into RL training — commission + spread = total tx cost

2026-03-01 22:10:43 +01:00

cuda_test.rs

fix: wire real GPU backprop + SPSA gradients, fix checkpoint loading, eliminate candle from examples

2026-03-20 15:41:42 +01:00

download_baseline.rs

fix: add retry with backoff to Databento download_quarter

2026-03-31 22:57:43 +02:00

evaluate_baseline.rs

fix: evaluate_baseline bf16→f32 remnants, delete old fxcache files

2026-04-10 18:57:36 +02:00

evaluate_supervised.rs

fix: wire real GPU backprop + SPSA gradients, fix checkpoint loading, eliminate candle from examples

2026-03-20 15:41:42 +01:00

hyperopt_baseline_rl.rs

feat: --feature-cache-dir CLI arg for training binaries

2026-04-03 22:59:08 +02:00

hyperopt_baseline_supervised.rs

fix(clippy): resolve all clippy warnings across entire workspace (--all-targets)

2026-03-13 10:18:35 +01:00

precompute_features.rs

refactor: fxcache single f32 format — delete bf16 version, remove --bf16 CLI flag

2026-04-10 18:45:49 +02:00

train_baseline_rl.rs

feat: leverage-based position cap (replaces hardcoded max_position)

2026-04-07 16:24:20 +02:00

train_baseline_supervised.rs

refactor: remove max_training_steps_per_epoch — always train full dataset

2026-04-02 14:41:16 +02:00