TensorImpl size constraints

Cloned pythourch from starts .CMake Build All in VS 2017
Everything else seems to be installed as by installation guide…

Failing on install Caffe 2…
Caffe 2 You changed the size of TensorImpl on 64-bit arch.See Note [TensorImpl size constraints] on how to proceed.

I have 16 GB RAM
Is it about memory?
How can I solve this?

at
D:\dev\ML\pytorch\aten\src\ATen\core\TensorImpl.h

line 1408
// Note [TensorImpl size constraints]
// ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
// Changed the size of TensorImpl? If the size went down, good for
// you! Adjust the documentation below and the expected size.
// Did it go up? Read on…
//
// Struct size matters. In some production systems at Facebook, we have
// 400M live tensors during a training run. Do the math: every 64-bit
// word you add to Tensor is an extra 3.2 gigabytes in RAM.
//
// If you are a Facebook employee, you can check if the run in question
// has tipped you over the point using the command here:
// https://fburl.com/q5enpv98
//
// For reference, we OOMed at 160 bytes (20 words) per TensorImpl.
// This is not counting overhead from strides out-of-line allocation, and
// StorageImpl space. We’re currently comfortably under this number;
// let’s keep it that way. (One currently approved pending size
// increase is inlining sizes and strides as small vectors, to reduce
// dynamic allocations.)
//
// Our memory usage on 32-bit systems is suboptimal, but we’re not checking
// for it at the moment (to help avoid rage inducing cycles when the
// 32-bit number is wrong).
//
// Current breakdown:
//
// vtable pointer
// strong refcount TODO: pack these into one word
// weak refcount
// storage pointer
// sizes vector (start)
// sizes vector (end)
// sizes vector (reserved) TODO: get rid of me
// strides pointer
// storage offset
// numel
// data type pointer
// miscellaneous bitfield
//
static_assert(sizeof(void*) != sizeof(int64_t) || // if 64-bit…
sizeof(TensorImpl) == sizeof(int64_t) * 12,
“You changed the size of TensorImpl on 64-bit arch.”
“See Note [TensorImpl size constraints] on how to proceed.”);

} // namespace at