TF Quant Finance — Derivatives Pricing Library

TF Quant Finance — High-Performance Derivatives Pricing Library

TF Quant Finance is Google's open-source TensorFlow-based library for quantitative finance. It provides GPU/TPU-accelerated implementations of mathematical finance models — Black-Scholes, Heston, SABR, HJM, Hull-White — for fast option pricing, risk sensitivities (Greeks), curve calibration, and Monte Carlo simulation at scale.

Key Features

• Option pricing: European, American, Asian, barrier, digital
• Rate models: Hull-White 1F/2F, HJM, LMM
• Volatility surface calibration (SABR, SVI)
• GPU-accelerated Monte Carlo simulation
• Automatic differentiation for Greeks via TensorFlow AD
• Numerical methods: PDE solvers, quadrature, root-finding

Quick Start

import tensorflow as tf
import tf_quant_finance as tff

# Black-Scholes European call price
price = tff.black_scholes.option_price(
    volatilities=tf.constant([0.2]),
    strikes=tf.constant([100.0]),
    expiries=tf.constant([1.0]),
    spots=tf.constant([100.0]),
    discount_factors=tf.constant([0.95]),
)
print(price.numpy())  # e.g. [8.916]

npx ai-supply add tf-quant-finance-derivatives-pricing

Curated mirror of the open-source TF Quant Finance (Apache-2.0). Get it from the source.