I’m trying to reproduce the PymC3 AR(1) tutorial in PyMC4. Here is what I am trying to do.
import pymc4 as pm import numpy as np import matplotlib.pyplot as plt import tensorflow as tf import arviz as az np.random.seed(seed=42) T = 100 y = np.zeros((T,)) for i in range(1,T): y[i] = 0.95 * y[i-1] + np.random.normal() y = y.reshape(len(y), -1) #plt.plot(y); @pm.model def model(data): x = yield pm.Normal(loc=0., scale=1., name="beta") like = yield pm.AR( name='like', num_timesteps=100, coefficients=[x], level_scale=0.5, observed_time_series=data) return like trace = pm.sample(model(y), num_samples=500)
However, it does not allow me to use
x as the coefficients and I’m getting the error:
TypeError: Value passed to parameter 'input' has DataType variant not in list of allowed values: float32, float64, int32, uint8, int16, int8, complex64, int64, qint8, quint8, qint32, bfloat16, uint16, complex128, float16, uint32, uint64
My aim is to add the AR(1) simulation example to pymc4/notebooks. Hope I have not missed any obvious guideline.