Frequently asked questions

“ImportError: dynamic module does not define init function”

  1. Make sure that the name specified in pybind::module and PYBIND11_PLUGIN is consistent and identical to the filename of the extension library. The latter should not contain any extra prefixes (e.g. instead of
  2. If the above did not fix your issue, then you are likely using an incompatible version of Python (for instance, the extension library was compiled against Python 2, while the interpreter is running on top of some version of Python 3, or vice versa)

“Symbol not found: __Py_ZeroStruct / _PyInstanceMethod_Type

See item 2 of the first answer.

“SystemError: dynamic module not initialized properly”

See item 2 of the first answer.

The Python interpreter immediately crashes when importing my module

See item 2 of the first answer.

CMake doesn’t detect the right Python version

The CMake-based build system will try to automatically detect the installed version of Python and link against that. When this fails, or when there are multiple versions of Python and it finds the wrong one, delete CMakeCache.txt and then invoke CMake as follows:

cmake -DPYTHON_EXECUTABLE:FILEPATH=<path-to-python-executable> .

Limitations involving reference arguments

In C++, it’s fairly common to pass arguments using mutable references or mutable pointers, which allows both read and write access to the value supplied by the caller. This is sometimes done for efficiency reasons, or to realize functions that have multiple return values. Here are two very basic examples:

void increment(int &i) { i++; }
void increment_ptr(int *i) { (*i)++; }

In Python, all arguments are passed by reference, so there is no general issue in binding such code from Python.

However, certain basic Python types (like str, int, bool, float, etc.) are immutable. This means that the following attempt to port the function to Python doesn’t have the same effect on the value provided by the caller – in fact, it does nothing at all.

def increment(i):
    i += 1 # nope..

pybind11 is also affected by such language-level conventions, which means that binding increment or increment_ptr will also create Python functions that don’t modify their arguments.

Although inconvenient, one workaround is to encapsulate the immutable types in a custom type that does allow modifications.

An other alternative involves binding a small wrapper lambda function that returns a tuple with all output arguments (see the remainder of the documentation for examples on binding lambda functions). An example:

int foo(int &i) { i++; return 123; }

and the binding code

m.def("foo", [](int i) { int rv = foo(i); return std::make_tuple(rv, i); });

How can I reduce the build time?

It’s good practice to split binding code over multiple files, as in the following example:


void init_ex1(py::module &);
void init_ex2(py::module &);
/* ... */

PYBIND11_PLUGIN(example) {
    py::module m("example", "pybind example plugin");

    /* ... */

    return m.ptr();


void init_ex1(py::module &m) {
    m.def("add", [](int a, int b) { return a + b; });


void init_ex1(py::module &m) {
    m.def("sub", [](int a, int b) { return a - b; });


>>> import example
>>> example.add(1, 2)
>>> example.sub(1, 1)

As shown above, the various init_ex functions should be contained in separate files that can be compiled independently from one another, and then linked together into the same final shared object. Following this approach will:

  1. reduce memory requirements per compilation unit.
  2. enable parallel builds (if desired).
  3. allow for faster incremental builds. For instance, when a single class definition is changed, only a subset of the binding code will generally need to be recompiled.

“recursive template instantiation exceeded maximum depth of 256”

If you receive an error about excessive recursive template evaluation, try specifying a larger value, e.g. -ftemplate-depth=1024 on GCC/Clang. The culprit is generally the generation of function signatures at compile time using C++14 template metaprogramming.

How can I create smaller binaries?

To do its job, pybind11 extensively relies on a programming technique known as template metaprogramming, which is a way of performing computation at compile time using type information. Template metaprogamming usually instantiates code involving significant numbers of deeply nested types that are either completely removed or reduced to just a few instructions during the compiler’s optimization phase. However, due to the nested nature of these types, the resulting symbol names in the compiled extension library can be extremely long. For instance, the included test suite contains the following symbol:


which is the mangled form of the following function type:

pybind11::cpp_function::cpp_function<void, Example2, std::__1::vector<std::__1::basic_string<wchar_t, std::__1::char_traits<wchar_t>, std::__1::allocator<wchar_t> >, std::__1::allocator<std::__1::basic_string<wchar_t, std::__1::char_traits<wchar_t>, std::__1::allocator<wchar_t> > > >&, pybind11::name, pybind11::sibling, pybind11::is_method, char [28]>(void (Example2::*)(std::__1::vector<std::__1::basic_string<wchar_t, std::__1::char_traits<wchar_t>, std::__1::allocator<wchar_t> >, std::__1::allocator<std::__1::basic_string<wchar_t, std::__1::char_traits<wchar_t>, std::__1::allocator<wchar_t> > > >&), pybind11::name const&, pybind11::sibling const&, pybind11::is_method const&, char const (&) [28])

The memory needed to store just the mangled name of this function (196 bytes) is larger than the actual piece of code (111 bytes) it represents! On the other hand, it’s silly to even give this function a name – after all, it’s just a tiny cog in a bigger piece of machinery that is not exposed to the outside world. So we’ll generally only want to export symbols for those functions which are actually called from the outside.

This can be achieved by specifying the parameter -fvisibility=hidden to GCC and Clang, which sets the default symbol visibility to hidden. It’s best to do this only for release builds, since the symbol names can be helpful in debugging sessions. On Visual Studio, symbols are already hidden by default, so nothing needs to be done there. Needless to say, this has a tremendous impact on the final binary size of the resulting extension library.

Another aspect that can require a fair bit of code are function signature descriptions. pybind11 automatically generates human-readable function signatures for docstrings, e.g.:

|  __init__(...)
|      __init__(*args, **kwargs)
|      Overloaded function.
|      1. __init__(example.Example1) -> NoneType
|      Docstring for overload #1 goes here
|      2. __init__(example.Example1, int) -> NoneType
|      Docstring for overload #2 goes here
|      3. __init__(example.Example1, example.Example1) -> NoneType
|      Docstring for overload #3 goes here

In C++11 mode, these are generated at run time using string concatenation, which can amount to 10-20% of the size of the resulting binary. If you can, enable C++14 language features (using -std=c++14 for GCC/Clang), in which case signatures are efficiently pre-generated at compile time. Unfortunately, Visual Studio’s C++14 support (constexpr) is not good enough as of April 2016, so it always uses the more expensive run-time approach.

Working with ancient Visual Studio 2009 builds on Windows

The official Windows distributions of Python are compiled using truly ancient versions of Visual Studio that lack good C++11 support. Some users implicitly assume that it would be impossible to load a plugin built with Visual Studio 2015 into a Python distribution that was compiled using Visual Studio 2009. However, no such issue exists: it’s perfectly legitimate to interface DLLs that are built with different compilers and/or C libraries. Common gotchas to watch out for involve not free()-ing memory region that that were malloc()-ed in another shared library, using data structures with incompatible ABIs, and so on. pybind11 is very careful not to make these types of mistakes.