ibtsam/nq-fvg-detector

NQ Fair Value Gap (FVG) Detector (try it out )

YOLOv26m model trained to detect Bullish and Bearish Fair Value Gaps (FVGs) on NQ Micro E-mini Nasdaq-100 Futures charts.

Test Set Results

Metric	Score
mAP@50	92.3%
mAP@50-95	58.2%
Precision	83.9%
Recall	89.4%
F1	86.6%

Usage

from ultralytics import YOLO

model = YOLO("ibtsam/nq-fvg-detector")
results = model("chart.png", conf=0.25)
results[0].show()

Classes

• bullish_fvg: gap created during upward impulse move
• bearish_fvg: gap created during downward impulse move

Dataset

nq-fvg-dataset on Kaggle

Details

• Instrument: NQ1! Futures
• Timeframes: 1m, 5m
• Input size: 640x640
• Framework: Ultralytics YOLOv26m

Limitations

Trained on white background TradingView charts only. Not financial advice.

Author

Ibtsam Sadiq — GitHub · LinkedIn