Reducing divergences : Implementation challenges

[pythonometrist]

@pythonometrist

M = X_train.shape[1]
    with pm.Model() as model:
         ν = 1
         λ = pm.HalfStudentT('lambda', 1)
        
        τ2 = pm.InverseGamma('tau_sq', alpha=0.5*ν, beta=0.5*ν/(λ), shape=M, testval=0.1)
        σ1 = pm.HalfNormal("σ1", 2.5)
        sd_ =pm.Deterministic('sd_', σ1 * np.sqrt(τ2))
        β = pm.Normal('beta', mu=0, sd=sd_, shape=M)
        σ = pm.HalfNormal("σ", 2.5)       
        β0 = pm.Normal("β0", 0, 10.)
        
        pred = pm.Data("pred", X_train)

        mu = pm.Deterministic('mu',β0 + pm.math.dot(pred, β))
        obs = pm.Normal("obs",mu , σ, observed=y_train)

Here is the updated version of the code.