gavindiaz/NexQuant
1
0
Fork 0
mirror of https://github.com/NicolasBohn/NexQuant.git synced 2026-07-27 23:47:46 +00:00
Code Issues Packages Projects Releases Wiki Activity
1,404 Commits 21 Branches 34 Tags
cbe1c52e0049bb56d2226215ac262f21b3ecd478
Commit Graph

9 Commits

Author SHA1 Message Date
TPTBusiness cbe1c52e00 refactor: rename project from Predix to NexQuant 
Rename all source files, scripts, tests, documentation, and configuration
from Predix/predix to NexQuant/nexquant across the entire codebase.
 2026-05-09 17:48:22 +02:00
TPTBusiness 0aea8c7671 fix: restore KronosPredictor instantiation deleted during refactor 2026-05-07 21:52:37 +02:00
TPTBusiness 72e8a4306e feat: support Kronos-small and Kronos-base models, auto-select GPU/CPU 2026-05-07 21:45:28 +02:00
TPTBusiness e0c287a575 feat: run Kronos on CPU to avoid GPU conflict with llama-server 2026-05-05 15:29:01 +02:00
TPTBusiness 01ba45be56 fix(kronos): replace rdagent_logger with stdlib logging for CI compatibility 2026-04-18 12:24:47 +02:00
TPTBusiness 888e841366 perf(kronos): batch GPU inference via predict_batch — 75x faster 
Replace sequential predict() calls with predict_batch() in both
build_kronos_factor and evaluate_kronos_model. Up to batch_size windows
processed simultaneously on GPU, reducing per-window time from ~10s to
~0.13s (10 windows in 1.3s on RTX 5060 Ti, 75x speedup).

Adds --batch-size / -b option (default 32) to both kronos-factor and
kronos-eval CLI commands. Falls back to single inference per window if a
batch fails. Refactors timestamp prep into _build_window_inputs helper.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-18 12:10:30 +02:00
TPTBusiness cffb9adc38 fix(kronos): pass actual datetime Series to Kronos predictor timestamps 
KronosPredictor.predict() requires x_timestamp and y_timestamp to be
pandas Series of datetime values for its calc_time_stamps() helper.
Previously we passed integer ranges (after reset_index), which raised
AttributeError on .dt.minute. Fixed by extracting datetime index values
before resetting and using future_idx for y_timestamp.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-18 10:12:51 +02:00
TPTBusiness 6e52c8a15d fix(kronos): lazy torch import to fix CI ModuleNotFoundError 
Move top-level `import torch` into _cuda_available() helper so
kronos_adapter.py can be imported in CI environments without torch.
All device defaults resolved at runtime via lazy detection.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-18 10:00:45 +02:00
TPTBusiness 1f6990d04d feat: integrate Kronos-mini OHLCV foundation model (Option A + B) 
Add Kronos-mini (4.1M params, AAAI 2026, MIT) as:
- Option A: predicted-return alpha factor via rolling daily inference
  (kronos_factor_gen.py — stride=96 bars/day, ~2k inference calls)
- Option B: standalone model evaluator alongside LightGBM
  (kronos_model_eval.py — IC / hit-rate vs actual realized returns)

KronosAdapter wraps NeoQuasar/Kronos-mini + Kronos-Tokenizer-2k,
auto-detects GPU, gracefully degrades if ~/Kronos repo is missing.
Factor output: MultiIndex (datetime, instrument) with KronosPredReturn.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
 2026-04-18 09:49:25 +02:00