replaced with an alternative API in `tf.contrib.seq2seq`.

*   Intel MKL Integration
    (https://software.intel.com/en-us/articles/tensorflow-optimizations-on-modern-intel-architecture).
    Intel developed a number of optimized deep learning primitives: In addition
    to matrix multiplication and convolution, these building blocks include:

curves from NIST SP 800-186. // // This package uses fiat-crypto or specialized assembly and Go code for its // backend field arithmetic (not math/big) and exposes constant-time, heap // allocation-free, byte slice-based safe APIs. Group operations use modern and // safe complete addition formulas where possible. The point at infinity is // handled and encoded according to SEC 1, Version 2.0, and invalid curve points // can't be represented. package nistec import _ "crypto/internal/fips140/check" //go:generate...