CmdStan User's Guide
Introduction
QuickStart Guide
1
CmdStan Installation
1.1
Installation via
conda
1.1.1
CmdStan install location under conda
1.2
Installation from GitHub
1.2.1
Downloading the source code
1.2.2
Building CmdStan
1.3
Checking the Stan compiler
1.4
Troubleshooting the installation
1.4.1
Common problems
1.5
C++ Toolchain
1.6
Using GNU Make
2
Example Model and Data
3
Compiling a Stan Program
3.1
Invoking the Make utility
3.2
Dependencies
3.3
Compiler errors
3.4
Troubleshooting C++ compiler or linker errors
3.5
C++ compilation and linking flags
3.5.1
Optimizing by ignoring range checks
4
MCMC Sampling
4.1
Running the sampler
4.2
Running multiple chains
4.2.1
Using the num_chains argument to run multiple chains
4.2.2
Using shell for running multiple chains
4.3
Stan CSV output file
4.4
Summarizing sampler output(s) with
stansummary
5
Optimization
6
Variational Inference
7
Generating Quantities of Interest from a Fitted Model
Reference Manual
8
Command-Line Interface Overview
8.1
Input data argument
8.2
Output control arguments
8.3
Initialize model parameters argument
8.4
Random number generator arguments
8.5
Chain identifier argument:
id
8.6
Command line help
8.7
Error messages and return codes
9
MCMC Sampling using Hamiltonian Monte Carlo
9.1
Iterations
9.2
Adaptation
9.2.1
Step size optimization configuration
9.2.2
Warmup schedule configuration
9.3
Algorithm
9.3.1
Samples from a set of fixed parameters
9.3.2
HMC samplers
9.4
Sampler diagnostic file
9.5
Examples
9.5.1
Running multiple chains with a specified RNG seed
9.5.2
Changing the default warmup and sampling iterations
9.5.3
Saving warmup draws
9.5.4
Initializing parameters
9.5.5
Specifying the metric and stepsize
9.5.6
Changing the NUTS-HMC adaptation parameters
9.5.7
Increasing the tree-depth
9.5.8
Capturing Hamiltonian diagnostics and gradients
9.5.9
Suppressing progress updates to the console
9.5.10
Everything example
10
Maximum Likelihood Estimation
10.1
Optimization algorithms
10.2
The quasi-Newton optimizers
10.3
The Newton optimizer
11
Variational Inference Algorithm: ADVI
11.1
Variational algorithms
11.2
Configuration
11.3
CSV output
12
Standalone Generate Quantities
13
Extracting log probabilities and gradients for diagnostics
13.1
Configuration
13.2
CSV output
14
Diagnosing HMC by Comparison of Gradients
15
Parallelization
15.1
Multi-threading with TBB
15.1.1
Compiling
15.1.2
Running
15.2
Multi-processing with MPI
15.2.1
Dependencies
15.2.2
Compiling
15.2.3
Running
15.3
OpenCL
15.3.1
Dependencies
15.3.2
Compiling
15.3.3
Running
CmdStan Tools
16
stanc
: Translating Stan to C++
16.1
Instantiating the
stanc
binary
16.2
The Stan compiler program
17
stansummary
: MCMC Output Analysis
17.1
Building the stansummary command
17.2
Running the
stansummary
program
17.2.1
Sampler parameters
17.2.2
Model parameters and quantities of interest
17.3
Command-line options
18
diagnose
: Diagnosing Biased Hamiltonian Monte Carlo Inferences
18.1
Building the diagnose command
18.2
Running the diagnose command
18.3
diagnose
warnings and recommendations
18.3.1
Divergent transitions after warmup
18.3.2
Maximum treedepth exceeded
18.3.3
Low E-BFMI values - sampler transitions HMC potential energy.
18.3.4
Low effective sample sizes
18.3.5
High
\(\hat{R}\)
19
print
(deprecated): MCMC Output Analysis
Appendices
20
Stan CSV File Format
20.1
CSV column names and order
20.2
MCMC sampler CSV output
20.2.1
Sampler Stan CSV output file
20.2.2
Diagnostic CSV output file
20.2.3
Profiling CSV output file
20.3
Optimization output
20.4
Variational inference output
20.5
Generate quantities outputs
20.6
Diagnose method outputs
21
JSON Format for CmdStan
21.1
Creating JSON files
21.2
JSON syntax summary
21.3
Stan data types in JSON notation
21.3.1
Empty arrays in JSON
22
RDump Format for CmdStan
22.1
Creating dump files
22.2
Scalar variables
22.3
Sequence variables
22.4
Array variables
22.5
Matrix- and vector-valued variables
22.5.1
Vector dump format
22.5.2
Matrix dump format
22.5.3
Arrays of vectors and matrices
22.6
Complex-valued variables
22.7
Integer- and real-valued variables
22.7.1
Scientific notation
22.7.2
Infinite and not-a-number values
22.8
Quoted variable names
22.9
Line breaks
22.10
BNF grammar for dump data
23
Using external C++ code
Bibliography
CmdStan User’s Guide
This is an old version,
view current version
.
CmdStan Tools
This section provides a reference for the CmdStan tools:
stanc
stansummary
diagnose
print (deprecated)