87b8303950
Per project_ml_alpha_starting_capital greenfield posture: there is no V1 to differentiate from (the V1 trunk forward was dead code, no V1 checkpoint files exist in the wild). The 'v2' prefix on every identifier was historical baggage from the migration period. Renames: - CheckpointV2 -> Checkpoint (also drops the version: u32 field — bincode either deserialises a current envelope or errors; no migration path needed) - CheckpointVersionProbe removed (was only for V1 rejection) - LAYER_NORM_CUBIN_V2 / VARIABLE_SELECTION_CUBIN_V2 / ATTENTION_POOL_CUBIN_V2 -> LAYER_NORM_CUBIN / VARIABLE_SELECTION_CUBIN / ATTENTION_POOL_CUBIN - _ln_module_v2 / _vsn_module_v2 / _attn_module_v2 -> drop _v2 suffix - smoke_load_v2_checkpoint test -> smoke_load_checkpoint - config/ml/sweep_v2_*.yaml -> config/ml/sweep_*.yaml - migration-era 'V2 weight skeleton' / 'V2 fields' / etc. comments cleaned to remove the v2 prefix Pre-existing 'v2' references in ml-backtesting CUDA files (decision_policy.cu, pnl_track.cu) are NOT touched — those refer to future planned 'v2' refinements (Portfolio mode, multi-fill averaging) from the C1-C19 commits and reflect aspirational features unrelated to this session's trunk-grows work. Verification: ml-alpha + ml-backtesting + fxt-backtest all build clean. perception_forward_golden bit-exact (max_diff = 0.000000).
16 lines
725 B
YAML
16 lines
725 B
YAML
# Spec §3.3 — threshold pre-registration on W0 only.
|
|
# Single-process fxt-backtest sweep (not Argo fan-out). 8 cells.
|
|
# Selects the production threshold by maximising in-sample Sharpe.
|
|
checkpoint: /feature-cache/alpha-perception-runs/__SHA__/trunk_best_h6000.bin
|
|
windows:
|
|
- id: W0
|
|
dbn_file: /data/futures-baseline-mbp10/ES.FUT/ES.FUT_2025-Q1.dbn.zst
|
|
axes:
|
|
cost_tick: [1.0] # anchor cost (1 tick all-in)
|
|
latency_ms: [200] # anchor latency (Scaleway PAR → IBKR realistic RTT)
|
|
threshold: [0.60, 0.65, 0.70, 0.75, 0.80, 0.85, 0.90, 0.95]
|
|
decision_stride: 1
|
|
horizons_used: [h6000]
|
|
# Selected threshold is persisted to config/ml/v2_prod_thresholds.json
|
|
# and committed before the deployability sweep runs.
|