TensorFloat-32 is enabled by default, but can be disabled with
        `tf.config.experimental.enable_tensor_float_32_execution`.
*   `tf.math`:
    *   Adds `tf.math.erfcinv`, the inverse to `tf.math.erfc`.
*   `tf.nn`:
    *   `tf.nn.max_pool2d` now supports explicit padding.
*   `tf.image`:
    *   Adds deterministic `tf.image.stateless_random_*` functions for each

nonce, ciphertext, data []byte) ([]byte, error) { fips140.RecordApproved() return g.g.Open(dst, nonce, ciphertext, data) } // NewGCMForTLS12 returns a new AEAD that works like GCM, but enforces the // construction of nonces as specified in RFC 5288, Section 3 and RFC 9325, // Section 7.2.1. // // This complies with FIPS 140-3 IG C.H Scenario 1.a. func NewGCMForTLS12(cipher *aes.Block) (*GCMForTLS12, error) { g, err := newGCM(&GCM{}, cipher, gcmStandardNonceSize, gcmTagSize) if err != nil { return nil,...