From 88b6efd0f2bb3e80de5c41ad0d53dc6a7cedbe5f Mon Sep 17 00:00:00 2001
From: Chuck Witt <witt.chuck@gmail.com>
Date: Tue, 31 Jan 2023 17:49:39 +0000
Subject: [PATCH] update tutorial.

---
 tutorials/first_example_basis.jl | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/tutorials/first_example_basis.jl b/tutorials/first_example_basis.jl
index df39fa9d..e0368e32 100644
--- a/tutorials/first_example_basis.jl
+++ b/tutorials/first_example_basis.jl
@@ -47,7 +47,8 @@ train = [gen_dat() for _=1:20];
 # We specify a solver and then as `ACEfit.jl` to do all the work for us. More fine-grained control is possible; see the `ACEfit.jl` documentation.
 # For sake of illustration we use a Bayesian Ridge Regression solver. This will automatically determine the regularisation for us. 
 
-solver = ACEfit.RRQR(rtol = 1e-4)   
+# TODO: change this solver name once we agree on the new API
+solver = ACEfit.BayesianLinearRegressionSVD() 
 solution = ACEfit.linear_fit(train, basis, solver)
 
 # Finally, we generate the potential from the parameters.