CALL p256MulInternal<>(SB) // s2 = y2 * z1ˆ3 STy(s2) LDx(y1in) CALL p256SubInternal<>(SB) // r = s2 - s1 STx(r) CALL p256SqrInternal<>(SB) // rsqr = rˆ2 STy (rsqr) LDx(h) CALL p256SqrInternal<>(SB) // hsqr = hˆ2 STy(hsqr) CALL p256MulInternal<>(SB) // hcub = hˆ3 STy(hcub) LDx(y1in) CALL p256MulInternal<>(SB) // y1 * hˆ3 STy(s2) LDP hsqr(0*8), (x0, x1) LDP hsqr(2*8), (x2, x3) LDP 0*16(a_ptr), (y0, y1) LDP 1*16(a_ptr), (y2, y3) CALL p256MulInternal<>(SB) // u1 * hˆ2 STP (y0, y1), h(0*8) STP (y2, y3),...

       *    `TfLiteSignatureRunnerResizeInputTensor`

    * New C API function `TfLiteExtensionApisVersion` added to `tensorflow/lite/c/c_api.h`.

    * Add int8 and int16x8 support for RSQRT operator

* Android NDK r25 is supported.

### Bug Fixes and Other Changes

*   Add TensorFlow Quantizer to TensorFlow pip package.