Value error when running the nascar example

[juliagorman]

I was just trying to run rSLDS on the NASCAR example by just running in a notebook cell
% run examples/nascar.py as instructed by the tutorial.

I then got this error:
ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (4,) + inhomogeneous part.

with this being the whole error message:

`Setting seed to 0

ValueError Traceback (most recent call last)
File /mnt/cube/jugorman/rSLDS/recurrent-slds/examples/nascar.py:581
576 plt.show()
579 if name == "main":
580 # Simulate NASCAR data
--> 581 true_model, inputs, z_true, x_true, y, mask = simulate_nascar()
583 # Run PCA to get 2D dynamics
584 x_init, C_init = fit_pca(y)

File /mnt/cube/jugorman/rSLDS/recurrent-slds/examples/nascar.py:136, in simulate_nascar()
133 reg_b += mu_b[:,None]
135 # Make a recurrent SLDS with these params #
--> 136 dynamics_distns = [
137 Regression(
138 A=np.column_stack((A,b)),
139 sigma=1e-4 * np.eye(D_latent),
140 nu_0=D_latent + 2,
141 S_0=1e-4 * np.eye(D_latent),
142 M_0=np.zeros((D_latent, D_latent + 1)),
143 K_0=np.eye(D_latent + 1),
144 )
145 for A,b in zip(As, bs)]
147 init_dynamics_distns = [
148 Gaussian(
149 mu=np.array([0.0, 1.0]),
150 sigma=1e-3 * np.eye(D_latent))
151 for _ in range(K_true)]
153 C = np.hstack((npr.randn(args.D_obs, D_latent), np.zeros((args.D_obs, 1))))

File /mnt/cube/jugorman/rSLDS/recurrent-slds/examples/nascar.py:137, in (.0)
133 reg_b += mu_b[:,None]
135 # Make a recurrent SLDS with these params #
136 dynamics_distns = [
--> 137 Regression(
138 A=np.column_stack((A,b)),
139 sigma=1e-4 * np.eye(D_latent),
140 nu_0=D_latent + 2,
141 S_0=1e-4 * np.eye(D_latent),
142 M_0=np.zeros((D_latent, D_latent + 1)),
143 K_0=np.eye(D_latent + 1),
144 )
145 for A,b in zip(As, bs)]
147 init_dynamics_distns = [
148 Gaussian(
149 mu=np.array([0.0, 1.0]),
150 sigma=1e-3 * np.eye(D_latent))
151 for _ in range(K_true)]
153 C = np.hstack((npr.randn(args.D_obs, D_latent), np.zeros((args.D_obs, 1))))

File /mnt/cube/jugorman/envs/rSLDS/lib/python3.8/site-packages/pybasicbayes/distributions/regression.py:36, in Regression.init(self, nu_0, S_0, M_0, K_0, affine, A, sigma)
32 have_hypers = not any_none(nu_0,S_0,M_0,K_0)
34 if have_hypers:
35 self.natural_hypparam = self.mf_natural_hypparam =
---> 36 self._standard_to_natural(nu_0,S_0,M_0,K_0)
38 if A is sigma is None and have_hypers:
39 self.resample()

File /mnt/cube/jugorman/envs/rSLDS/lib/python3.8/site-packages/pybasicbayes/distributions/regression.py:85, in Regression._standard_to_natural(nu, S, M, K)
83 C = Kinv
84 d = nu
---> 85 return np.array([A,B,C,d])`

All I have done is just downloaded the required packages and tried to run the example as it currently is