Question

当尝试适合简单模型时，我在控制台中收到此错误消息：

Chain 3, Iteration:  800 / 2000 [ 40%]  (Warmup)
Chain 2, Iteration:  800 / 2000 [ 40%]  (Warmup)
Chain 3, Iteration: 1000 / 2000 [ 50%]  (Warmup)python(9919,0x700001b19000) malloc: *** error for object 0x10b7c86e0: pointer being freed was not allocated
*** set a breakpoint in malloc_error_break to debug

Chain 2, Iteration: 1000 / 2000 [ 50%]  (Warmup)python(9918,0x700001b19000) malloc: *** error for object 0x10b7c86e0: pointer being freed was not allocated
*** set a breakpoint in malloc_error_break to debug

尝试用conda和/或pip安装pystan。

Python 3.5

OS X El capitan 10.11.4

尝试过Pystan 2.8.0.0和2.9.0.0

这是一个完整的代码，从Jupyter

复制粘贴

import numpy as np
import pystan

true_a = 5.6
true_b = 0.17
true_sigma = 3.0

np.random.seed(1)
x = np.random.uniform(0, 100, 200)
y = true_a + true_b*x + np.random.normal(0, true_sigma, len(x))

stan_datadict = {}
stan_datadict['N'] = len(x)
stan_datadict['x'] = x
stan_datadict['y'] = y

stan_ols = """

    data {                                 
        int<lower=0> N;               // count of observations
        vector[N] x;                  // exog features
        vector[N] y;                  // endog feature
    }
    parameters {
        real alpha;                   // intercept (default uniform [-inf, inf])
        real beta;                    // slope (default uniform [-inf, inf])
        real<lower=0> sigma;          // model error (default uniform [0, inf])
    }
    model {  
        y ~ normal(alpha + beta * x, sigma);    // normal likelihood
    }
    """
stan_fit_ols = pystan.stan(model_code=stan_ols, data=stan_datadict,
                           iter=2000, warmup=1000, chains=4, n_jobs=2,verbose=False)

pystan OS X El capitan malloc：***对象0x117e956e0的错误

0 个答案: