Bug: NaN in ADVI #1610

cemoody · 2015-09-09T17:38:43Z

Bug Report

Version

Using cmdstan-2.7.0 with Stan 2.7.0.

Description

Using the following Stan file with ADVI produces a NaN after 50 iterations.

data {
  int<lower=2> K;                // num topics
  int<lower=2> V;                // num words
  int<lower=2> U;                // num users
  int<lower=2> I;                // num items
  int<lower=1> N;                // total word instances
  int<lower=1,upper=V> word[N];  // word n
  int<lower=1,upper=I> item[N];  // item ID for word n
  int<lower=1,upper=U> user[N];  // user ID for word n
  vector<lower=0, upper=1>[K] alpha_user; // topic prior concentrations for users
  vector<lower=0, upper=1>[K] alpha_item; // topic prior concentrations for items
  vector<lower=0, upper=1>[V] beta;       // prior probability for seeing word
}
parameters {
  simplex[K] item_topics[I];    // topic dist for item i
  simplex[K] user_topics[U];    // topic dist for user u
  simplex[V] word_topics[K];    // for topic k prob of seeing word v
}
model {
  for (i in 1:I)  
    item_topics[i] ~ dirichlet(alpha_item);  // prior on item topics
  for (u in 1:U)  
    user_topics[u] ~ dirichlet(alpha_user);  // prior on user topics
  for (k in 1:K)  
    word_topics[k] ~ dirichlet(beta);     // prior
  // for every word in our corpus
  for (n in 1:N) {
    real gamma[K];
    // for every topic 
    for (k in 1:K){ 
      // topic distribution for this user
      gamma[k] <- log(item_topics[item[n], k] + user_topics[user[n], k]) + log(word_topics[k, word[n]]);
    }
    increment_log_prob(log_sum_exp(gamma));  // likelihood
  }
}

Steps

Compile Stan code:

~/cmdstan-2.7.0/bin/stanc lda_cf_bias.stan

Compile again:

g++ -I /home/moody/code/cmdstan-2.7.0/src -I /home/moody/code/cmdstan-2.7.0/stan/src -isystem /home/moody/code/cmdstan-2.7.0/stan/lib/stan_math_2.7.0 -isystem /home/moody/code/cmdstan-2.7.0/stan/lib/eigen_3.2.4 -isystem /home/moody/code/cmdstan-2.7.0/stan/lib/boost_1.58.0 -Wall -pipe -DEIGEN_NO_DEBUG -Wno-unused-local-typedefs -lpthread -O3 -o lda_cf_bias /home/moody/code/cmdstan-2.7.0/src/cmdstan/main.cpp -include lda_cf_bias_model.cpp

Run:

./lda_cf_bias variational elbo_samples=10 eval_elbo=10 iter=100 tol_rel_obj=0.001 random seed=3125502058 data file=lda_cf.data

Output

./lda_cf_bias variational elbo_samples=10 eval_elbo=10 iter=100 tol_rel_obj=0.001 data file=lda_cf.data
 method = variational
   variational
     algorithm = meanfield (Default)
       meanfield
     iter = 100
     grad_samples = 1 (Default)
     elbo_samples = 10
     eta_adagrad = 0.10000000000000001 (Default)
     tol_rel_obj = 0.001
     eval_elbo = 10
     output_samples = 1000 (Default)
 id = 0 (Default)
 data
   file = lda_cf.data
 init = 2 (Default)
 random
   seed = 3125502058
 output
   file = output.csv (Default)
   diagnostic_file =  (Default)
   refresh = 100 (Default)


This is Automatic Differentiation Variational Inference.

(EXPERIMENTAL ALGORITHM: expect frequent updates to the procedure.)

Gradient evaluation took 0.654544 seconds
1000 iterations under these settings should take 654.544 seconds.
Adjust your expectations accordingly!

  iter       ELBO   delta_ELBO_mean   delta_ELBO_med   notes
     0     -1e+06             1.000            1.000
    10  -1079912.8             0.137            0.137
    20  -1039906.1             0.038            0.038
    30  -1022367.2             0.017            0.017
    40  -996845.3             0.026            0.026
    50       -nan               nan              nan

Exception: stan::variational::normal_meanfield::set_mu: Input vector[55629] is -nan, but must not be nan!
Diagnostic information:
Dynamic exception type: std::domain_error
std::exception::what: stan::variational::normal_meanfield::set_mu: Input vector[55629] is -nan, but must not be nan!

Expected Output

Additional Information

Data used (e.g. lda_cf.data) is:

alpha_user <-
c(0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1)

alpha_item <-
c(0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1)

I <- 84

K <- 10

N <- 326

item <-
c(11, 11, 11, 11, 31, 62, 58, 58, 58, 58, 58, 58, 58, 52, 52, 52, 54, 54, 54, 57, 57, 35, 35, 35, 35, 35, 29, 29, 29, 80, 80, 80, 80, 81, 81, 78, 78, 78, 78, 78, 78, 48, 48, 48, 48, 48, 48, 48, 12, 12, 12, 5, 5, 5, 5, 5, 43, 43, 6, 6, 6, 74, 74, 74, 74, 74, 74, 74, 74, 74, 9, 9, 9, 9, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 76, 56, 56, 56, 56, 56, 56, 7, 41, 41, 41, 16, 16, 16, 16, 16, 67, 82, 82, 82, 82, 66, 66, 66, 64, 1, 1, 1, 1, 73, 60, 60, 60, 47, 47, 47, 25, 25, 25, 25, 38, 38, 38, 38, 53, 53, 53, 53, 15, 15, 15, 51, 51, 51, 51, 51, 62, 62, 46, 40, 27, 27, 27, 39, 64, 64, 64, 69, 69, 44, 55, 55, 55, 55, 55, 55, 55, 75, 75, 75, 75, 75, 49, 36, 36, 36, 36, 36, 36, 36, 36, 36, 22, 79, 79, 79, 59, 59, 30, 30, 67, 67, 67, 67, 67, 67, 37, 37, 37, 37, 37, 64, 8, 8, 32, 32, 32, 32, 32, 81, 81, 81, 81, 81, 20, 20, 20, 20, 20, 84, 33, 33, 10, 34, 34, 26, 2, 2, 2, 2, 68, 68, 68, 68, 68, 68, 68, 68, 68, 77, 77, 3, 3, 3, 3, 70, 70, 63, 63, 63, 14, 61, 23, 23, 23, 65, 45, 45, 45, 71, 71, 71, 71, 24, 24, 24, 24, 4, 4, 59, 59, 59, 59, 59, 63, 63, 59, 59, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 18, 18, 18, 18, 21, 21, 21, 21, 21, 21, 21, 21, 39, 39, 39, 39, 39, 39, 39, 17, 17, 17, 13, 13, 13, 42, 42, 42, 42, 42, 42, 83, 83, 19, 72)

beta <-
c(0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235)

U <- 28

user <-
c(11, 11, 11, 11, 7, 1, 4, 4, 4, 4, 4, 4, 4, 7, 7, 7, 5, 5, 5, 2, 2, 6, 6, 6, 6, 6, 12, 12, 12, 15, 15, 15, 15, 10, 10, 4, 4, 4, 4, 4, 4, 14, 14, 14, 14, 14, 14, 14, 12, 12, 12, 3, 3, 3, 3, 3, 13, 13, 8, 8, 8, 12, 12, 12, 12, 12, 12, 12, 12, 12, 3, 3, 3, 3, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 6, 10, 10, 10, 10, 10, 10, 4, 16, 16, 16, 13, 13, 13, 13, 13, 10, 15, 15, 15, 15, 2, 2, 2, 12, 8, 8, 8, 8, 1, 5, 5, 5, 11, 11, 11, 7, 7, 7, 7, 7, 7, 7, 7, 15, 15, 15, 15, 5, 5, 5, 3, 3, 3, 3, 3, 6, 6, 8, 12, 15, 15, 15, 16, 3, 3, 3, 9, 9, 8, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 6, 10, 10, 10, 10, 10, 10, 10, 10, 10, 6, 11, 11, 11, 15, 15, 1, 1, 9, 9, 9, 9, 9, 9, 10, 10, 10, 10, 10, 4, 21, 21, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 21, 21, 23, 12, 12, 12, 12, 12, 12, 19, 19, 19, 19, 26, 26, 26, 26, 26, 26, 26, 26, 26, 17, 17, 18, 18, 18, 18, 18, 18, 17, 17, 17, 24, 18, 17, 17, 17, 17, 18, 18, 18, 26, 26, 26, 26, 26, 26, 26, 26, 25, 25, 25, 25, 25, 25, 25, 25, 25, 26, 26, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 27, 27, 27, 27)

V <- 17

word <-
c(1, 2, 3, 4, 5, 6, 6, 7, 8, 9, 1, 10, 7, 6, 7, 8, 11, 12, 3, 11, 13, 14, 3, 5, 10, 10, 14, 4, 15, 16, 4, 5, 4, 2, 4, 13, 3, 14, 9, 3, 3, 1, 12, 10, 9, 11, 16, 12, 5, 4, 15, 13, 11, 9, 12, 8, 9, 5, 3, 5, 3, 1, 12, 10, 17, 13, 10, 10, 4, 15, 13, 9, 11, 14, 1, 2, 4, 15, 1, 13, 10, 4, 15, 9, 3, 5, 17, 1, 11, 10, 3, 17, 11, 12, 8, 4, 14, 4, 5, 8, 6, 17, 1, 12, 10, 9, 10, 4, 12, 3, 5, 4, 10, 5, 9, 13, 13, 10, 4, 15, 6, 13, 13, 14, 6, 7, 8, 6, 7, 7, 7, 6, 7, 7, 7, 13, 3, 17, 4, 11, 13, 3, 13, 11, 2, 12, 10, 13, 3, 10, 13, 4, 17, 4, 6, 6, 7, 8, 13, 3, 16, 13, 3, 9, 1, 11, 12, 3, 9, 1, 11, 12, 3, 17, 13, 9, 4, 10, 14, 1, 2, 3, 4, 5, 6, 16, 4, 11, 4, 13, 3, 11, 12, 9, 1, 12, 10, 13, 9, 5, 16, 4, 12, 6, 10, 6, 10, 8, 14, 4, 6, 10, 2, 10, 10, 6, 10, 3, 6, 10, 17, 14, 6, 10, 14, 16, 13, 5, 3, 13, 3, 1, 13, 3, 8, 9, 5, 17, 3, 6, 9, 5, 13, 9, 11, 10, 9, 16, 6, 7, 8, 5, 2, 6, 7, 8, 7, 6, 4, 15, 9, 1, 2, 16, 13, 8, 9, 16, 1, 11, 2, 16, 17, 4, 15, 13, 10, 9, 5, 13, 3, 17, 3, 14, 6, 3, 9, 2, 10, 10, 16, 2, 16, 4, 15, 6, 7, 8, 4, 15, 5, 4, 15, 13, 10, 9, 5, 17, 4, 15, 5, 4, 15, 5, 4, 15, 13, 10, 9, 5, 4, 15, 9, 11, 14, 13)

Running the test gradients seems fine:

./lda_cf_bias diagnose test=gradient random seed=3125502058 data file=lda_cf.data

./lda_cf_bias diagnose test=gradient random seed=3125502058 data
file=lda_cf.data
 method = diagnose
   diagnose
     test = gradient (Default)
       gradient
         epsilon = 9.9999999999999995e-07 (Default)
         error = 9.9999999999999995e-07 (Default)
 id = 0 (Default)
 data
   file = lda_cf.data
 init = 2 (Default)
 random
   seed = 3125502058
 output
   file = output.csv (Default)
   diagnostic_file =  (Default)
   refresh = 100 (Default)


TEST GRADIENT MODE

 Log probability=-1099.86

 param idx           value           model     finite diff           error
         0        -1.64857        0.055292       0.0552918     1.29484e-07
         1       -0.148741       0.0122065       0.0122063     1.50648e-07
         2        -1.04209        0.143936        0.143936     1.35776e-07
         3       -0.394122      0.00999742      0.00999762    -2.03224e-07
         4         1.57506       -0.136871        -0.13687    -4.85612e-07
         5       -0.983549       0.0399101       0.0399104    -3.13723e-07
         6         -0.3114      -0.0372367       -0.037237     2.86941e-07
         7       -0.620216       0.0319952       0.0319951     1.32904e-07
         8         1.63741     -0.00928295     -0.00928321     2.60286e-07
         9       0.0250666      -0.0502523      -0.0502519     -4.5979e-07
        10         1.13795       -0.120103       -0.120103     4.93569e-07
        11         -1.7728       0.0642492        0.064249     2.02201e-07
... [All of the errors are ~1e-7 or ~1e-08]...

The text was updated successfully, but these errors were encountered:

bob-carpenter · 2015-09-09T18:35:51Z

Thanks for reporting. This may be fixed with the new
adaptation and resampling that should go in Stan 2.9.

Bob

On Sep 9, 2015, at 1:38 PM, Christopher Erick Moody notifications@github.com wrote:

Bug Report

Version

Using cmdstan-2.7.0 with Stan 2.7.0.

Description

Using the following Stan file with ADVI produces a NaN after 50 iterations.

data {
int<lower=2> K; // num topics
int<lower=2> V; // num words
int<lower=2> U; // num users
int<lower=2> I; // num items
int<lower=1> N; // total word instances
int<lower=1,upper=V> word[N]; // word n
int<lower=1,upper=I> item[N]; // item ID for word n
int<lower=1,upper=U> user[N]; // user ID for word n
vector<lower=0, upper=1>[K] alpha_user; // topic prior concentrations for users
vector<lower=0, upper=1>[K] alpha_item; // topic prior concentrations for items
vector<lower=0, upper=1>[V] beta; // prior probability for seeing word
}
parameters {
simplex[K] item_topics[I]; // topic dist for item i
simplex[K] user_topics[U]; // topic dist for user u
simplex[V] word_topics[K]; // for topic k prob of seeing word v
}
model {
for (i in 1:I)
item_topics[i] ~ dirichlet(alpha_item); // prior on item topics
for (u in 1:U)
user_topics[u] ~ dirichlet(alpha_user); // prior on user topics
for (k in 1:K)
word_topics[k] ~ dirichlet(beta); // prior
// for every word in our corpus
for (n in 1:N) {
real gamma[K];
// for every topic
for (k in 1:K){
// topic distribution for this user
gamma[k] <- log(item_topics[item[n], k] + user_topics[user[n], k]) + log(word_topics[k, word[n]]);
}
increment_log_prob(log_sum_exp(gamma)); // likelihood
}
}

Steps

Compile Stan code:

~/cmdstan-2.7.0/bin/stanc lda_cf_bias.stan

Compile again:

g++ -I /home/moody/code/cmdstan-2.7.0/src -I /home/moody/code/cmdstan-2.7.0/stan/src -isystem /home/moody/code/cmdstan-2.7.0/stan/lib/stan_math_2.7.0 -isystem /home/moody/code/cmdstan-2.7.0/stan/lib/eigen_3.2.4 -isystem /home/moody/code/cmdstan-2.7.0/stan/lib/boost_1.58.0 -Wall -pipe -DEIGEN_NO_DEBUG -Wno-unused-local-typedefs -lpthread -O3 -o lda_cf_bias /home/moody/code/cmdstan-2.7.0/src/cmdstan/main.cpp -include lda_cf_bias_model.cpp

Run:

./lda_cf_bias variational elbo_samples=10 eval_elbo=10 iter=100 tol_rel_obj=0.001 random seed=3125502058 data file=lda_cf.data

Output

./lda_cf_bias variational elbo_samples=10 eval_elbo=10 iter=100 tol_rel_obj=0.001 data file=lda_cf.data
method = variational
variational
algorithm = meanfield (Default)
meanfield
iter = 100
grad_samples = 1 (Default)
elbo_samples = 10
eta_adagrad = 0.10000000000000001 (Default)
tol_rel_obj = 0.001
eval_elbo = 10
output_samples = 1000 (Default)
id = 0 (Default)
data
file = lda_cf.data
init = 2 (Default)
random
seed = 3125502058
output
file = output.csv (Default)
diagnostic_file = (Default)
refresh = 100 (Default)

This is Automatic Differentiation Variational Inference.

(EXPERIMENTAL ALGORITHM: expect frequent updates to the procedure.)

Gradient evaluation took 0.654544 seconds
1000 iterations under these settings should take 654.544 seconds.
Adjust your expectations accordingly!

iter ELBO delta_ELBO_mean delta_ELBO_med notes
0 -1e+06 1.000 1.000
10 -1079912.8 0.137 0.137
20 -1039906.1 0.038 0.038
30 -1022367.2 0.017 0.017
40 -996845.3 0.026 0.026
50 -nan nan nan

Exception: stan::variational::normal_meanfield::set_mu: Input vector[55629] is -nan, but must not be nan!
Diagnostic information:
Dynamic exception type: std::domain_error
std::exception::what: stan::variational::normal_meanfield::set_mu: Input vector[55629] is -nan, but must not be nan!

Expected Output

Additional Information

Data used (e.g. lda_cf.data) is:

alpha_user <-
c(0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1)

alpha_item <-
c(0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1)

I <- 84

K <- 10

N <- 326

item <-
c(11, 11, 11, 11, 31, 62, 58, 58, 58, 58, 58, 58, 58, 52, 52, 52, 54, 54, 54, 57, 57, 35, 35, 35, 35, 35, 29, 29, 29, 80, 80, 80, 80, 81, 81, 78, 78, 78, 78, 78, 78, 48, 48, 48, 48, 48, 48, 48, 12, 12, 12, 5, 5, 5, 5, 5, 43, 43, 6, 6, 6, 74, 74, 74, 74, 74, 74, 74, 74, 74, 9, 9, 9, 9, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 76, 56, 56, 56, 56, 56, 56, 7, 41, 41, 41, 16, 16, 16, 16, 16, 67, 82, 82, 82, 82, 66, 66, 66, 64, 1, 1, 1, 1, 73, 60, 60, 60, 47, 47, 47, 25, 25, 25, 25, 38, 38, 38, 38, 53, 53, 53, 53, 15, 15, 15, 51, 51, 51, 51, 51, 62, 62, 46, 40, 27, 27, 27, 39, 64, 64, 64, 69, 69, 44, 55, 55, 55, 55, 55, 55, 55, 75, 75, 75, 75, 75, 49, 36, 36, 36, 36, 36, 36, 36, 36, 36, 22, 79, 79, 79, 59, 59, 30, 30, 67, 67, 67, 67, 67, 67, 37, 37, 37, 37, 37, 64, 8, 8, 32, 32, 32, 32, 32, 81, 81, 81, 81, 81, 20, 20, 20, 20, 20, 84, 33, 33, 10, 34, 34, 26, 2, 2, 2, 2, 68, 68, 68, 68, 68, 68, 68, 68, 68, 77, 77, 3, 3, 3, 3, 70, 70, 63, 63, 63, 14, 61, 23,
23, 23,
65, 45, 45, 45, 71, 71, 71, 71, 24, 24, 24, 24, 4, 4, 59, 59, 59, 59, 59, 63, 63, 59, 59, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 18, 18, 18, 18, 21, 21, 21, 21, 21, 21, 21, 21, 39, 39, 39, 39, 39, 39, 39, 17, 17, 17, 13, 13, 13, 42, 42, 42, 42, 42, 42, 83, 83, 19, 72)

beta <-
c(0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235, 0.0588235)

U <- 28

user <-
c(11, 11, 11, 11, 7, 1, 4, 4, 4, 4, 4, 4, 4, 7, 7, 7, 5, 5, 5, 2, 2, 6, 6, 6, 6, 6, 12, 12, 12, 15, 15, 15, 15, 10, 10, 4, 4, 4, 4, 4, 4, 14, 14, 14, 14, 14, 14, 14, 12, 12, 12, 3, 3, 3, 3, 3, 13, 13, 8, 8, 8, 12, 12, 12, 12, 12, 12, 12, 12, 12, 3, 3, 3, 3, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 6, 10, 10, 10, 10, 10, 10, 4, 16, 16, 16, 13, 13, 13, 13, 13, 10, 15, 15, 15, 15, 2, 2, 2, 12, 8, 8, 8, 8, 1, 5, 5, 5, 11, 11, 11, 7, 7, 7, 7, 7, 7, 7, 7, 15, 15, 15, 15, 5, 5, 5, 3, 3, 3, 3, 3, 6, 6, 8, 12, 15, 15, 15, 16, 3, 3, 3, 9, 9, 8, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 6, 10, 10, 10, 10, 10, 10, 10, 10, 10, 6, 11, 11, 11, 15, 15, 1, 1, 9, 9, 9, 9, 9, 9, 10, 10, 10, 10, 10, 4, 21, 21, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 21, 21, 23, 12, 12, 12, 12, 12, 12, 19, 19, 19, 19, 26, 26, 26, 26, 26, 26, 26, 26, 26, 17, 17, 18, 18, 18, 18, 18, 18, 17, 17, 17, 24, 18, 17, 17, 17, 17, 18, 18, 18, 26, 26, 26, 26, 26, 26, 26, 26, 25, 25, 25, 25, 25, 25, 25, 25, 25
, 26, 26
, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 27, 27, 27, 27)

V <- 17

word <-
c(1, 2, 3, 4, 5, 6, 6, 7, 8, 9, 1, 10, 7, 6, 7, 8, 11, 12, 3, 11, 13, 14, 3, 5, 10, 10, 14, 4, 15, 16, 4, 5, 4, 2, 4, 13, 3, 14, 9, 3, 3, 1, 12, 10, 9, 11, 16, 12, 5, 4, 15, 13, 11, 9, 12, 8, 9, 5, 3, 5, 3, 1, 12, 10, 17, 13, 10, 10, 4, 15, 13, 9, 11, 14, 1, 2, 4, 15, 1, 13, 10, 4, 15, 9, 3, 5, 17, 1, 11, 10, 3, 17, 11, 12, 8, 4, 14, 4, 5, 8, 6, 17, 1, 12, 10, 9, 10, 4, 12, 3, 5, 4, 10, 5, 9, 13, 13, 10, 4, 15, 6, 13, 13, 14, 6, 7, 8, 6, 7, 7, 7, 6, 7, 7, 7, 13, 3, 17, 4, 11, 13, 3, 13, 11, 2, 12, 10, 13, 3, 10, 13, 4, 17, 4, 6, 6, 7, 8, 13, 3, 16, 13, 3, 9, 1, 11, 12, 3, 9, 1, 11, 12, 3, 17, 13, 9, 4, 10, 14, 1, 2, 3, 4, 5, 6, 16, 4, 11, 4, 13, 3, 11, 12, 9, 1, 12, 10, 13, 9, 5, 16, 4, 12, 6, 10, 6, 10, 8, 14, 4, 6, 10, 2, 10, 10, 6, 10, 3, 6, 10, 17, 14, 6, 10, 14, 16, 13, 5, 3, 13, 3, 1, 13, 3, 8, 9, 5, 17, 3, 6, 9, 5, 13, 9, 11, 10, 9, 16, 6, 7, 8, 5, 2, 6, 7, 8, 7, 6, 4, 15, 9, 1, 2, 16, 13, 8, 9, 16, 1, 11, 2, 16, 17, 4, 15, 13, 10, 9, 5, 13, 3, 17, 3, 14, 6, 3, 9, 2, 1
0, 10, 1
6, 2, 16, 4, 15, 6, 7, 8, 4, 15, 5, 4, 15, 13, 10, 9, 5, 17, 4, 15, 5, 4, 15, 5, 4, 15, 13, 10, 9, 5, 4, 15, 9, 11, 14, 13)

—
Reply to this email directly or view it on GitHub.

akucukelbir · 2015-11-23T21:27:05Z

@cemoody Fixed with #1699. Do let us know how the accuracy of the fits are! Cheers. Alp

syclik added this to the v2.8.0++ milestone Sep 17, 2015

syclik added the bug label Sep 17, 2015

syclik assigned akucukelbir Sep 17, 2015

dustinvtran assigned dustinvtran and unassigned akucukelbir Sep 22, 2015

akucukelbir closed this as completed Nov 23, 2015

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Bug: NaN in ADVI #1610

Bug: NaN in ADVI #1610

cemoody commented Sep 9, 2015

bob-carpenter commented Sep 9, 2015

akucukelbir commented Nov 23, 2015

Bug: NaN in ADVI #1610

Bug: NaN in ADVI #1610

Comments

cemoody commented Sep 9, 2015

Bug Report

bob-carpenter commented Sep 9, 2015

akucukelbir commented Nov 23, 2015