tf-quant-finance

What it is

A Python library built on TensorFlow that provides hardware acceleration and automatic differentiation for quantitative finance workloads. It is organized in three tiers: foundational mathematical methods such as optimisation, interpolation, root finders and random number generation; mid-level methods including ODE and PDE solvers, Ito process frameworks and copula samplers; and pricing models such as Local Vol, Stochastic Vol, Stochastic Local Vol and Hull-White, with calibration, rate curve building, payoff descriptions and schedule generation. Quantitative developers comfortable with TensorFlow will find it most useful, and self-study notebooks are included for those new to the framework. The project has been archived and is no longer maintained, so users are advised to fork it for continued development.