Mention Diffrax

README.md

@@ -1,6 +1,8 @@
 <h1 align='center'>torchcde</h1>
 <h2 align='center'>Differentiable GPU-capable solvers for CDEs</h2>
 
+**Update: for any new projects, I would now recommend using [Diffrax](https://github.com/patrick-kidger/diffrax) instead. This is much faster, and producion-quality. torchcde was its prototype as a research project!**
+
 This library provides differentiable GPU-capable solvers for controlled differential equations (CDEs). Backpropagation through the solver or via the adjoint method is supported; the latter allows for improved memory efficiency.
 
 In particular this allows for building [Neural Controlled Differential Equation](https://github.com/patrick-kidger/NeuralCDE) models, which are state-of-the-art models for (arbitrarily irregular!) time series. Neural CDEs can be thought of as a "continuous time RNN".