-
-
Notifications
You must be signed in to change notification settings - Fork 367
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Bug: NaN in ADVI #1610
Comments
Thanks for reporting. This may be fixed with the new
|
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.
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
Expected Output
Additional Information
Data used (e.g.
lda_cf.data
) is:Running the test gradients seems fine:
./lda_cf_bias diagnose test=gradient random seed=3125502058 data file=lda_cf.data
The text was updated successfully, but these errors were encountered: