Stage 1

ML Training

Dataset Training Mode

Collect feature snapshots, label future movement, and train an offline baseline model.
idle
Last runner error: kucoinfutures fetchOrderBook() limit argument must be 20 or 100

Dataset Stats

Current feature and label inventory.
Total feature snapshots
0
Total labels
0
Pending labels
0
Eligible pending labels
0
Labeled snapshots
0
Labels with advanced TP/SL
0
Incomplete advanced labels
0
Avg long MFE 30s
-
Avg long MAE 30s
-
Avg short MFE 30s
-
Avg short MAE 30s
-
Long labels
0
Short labels
0
Flat labels
0
Latest snapshot
No snapshots yet
Recent trades snapshots
0
Micro candles
0
Avg data quality
-
Latest latency
-
Missing fields
-

Active Exchanges / Pairs

Pairs currently eligible for ML feature collection.
Exchange Code Symbol
Binance binance VELVET/USDT:USDT
Gate.io gate VELVET/USDT:USDT
Huobi / HTX htx VELVET/USDT:USDT
KuCoin Futures kucoinfutures VELVET/USDT:USDT
MEXC mexc VELVET/USDT:USDT

Recent Feature Snapshots

Latest ML rows with tape pressure, candle momentum, deltas, and quality.
Page 1 of 1 · 0 total records · 10 per page

No feature snapshots yet

Start training collection to capture order book, tape, candle, funding, and quality features.

Snapshot Breakdown

Collected feature rows by exchange and symbol.

By Exchange

No exchange snapshots yet.

By Symbol

No symbol snapshots yet.

Trained Models

Offline baseline models trained from labeled historical data.

No trained models

Collect and label at least 1000 rows for a horizon before training a basic model.