IN1, H2H, XH1 // H^2.hi·H^2.hi VPMSUMD XL, XC2, T2 // 1st reduction phase VPMSUMD XL1, XC2, HH // 1st reduction phase VSLDOI $8, XM, ZERO, T0 VSLDOI $8, ZERO, XM, T1 VSLDOI $8, XM1, ZERO, HL VSLDOI $8, ZERO, XM1, H VXOR XL, T0, XL VXOR XH, T1, XH VXOR XL1, HL, XL1 VXOR XH1, H, XH1 VSLDOI $8, XL, XL, XL VSLDOI $8, XL1, XL1, XL1 VXOR XL, T2, XL VXOR XL1, HH, XL1 VSLDOI $8, XL, XL, T1 // 2nd reduction phase VSLDOI $8, XL1, XL1, H // 2nd reduction phase VPMSUMD XL, XC2, XL VPMSUMD XL1, XC2, XL1 VXOR T1,...

    *   New loss reduction types:
    *   `AUTO`: Indicates that the reduction option will be determined by the
        usage context. For almost all cases this defaults to
        `SUM_OVER_BATCH_SIZE`. When used with `tf.distribute.Strategy`, outside
        of built-in training loops such as `tf.keras` `compile` and `fit`, we
        expect reduction value to be `SUM` or `NONE`. Using `AUTO` in that case

     * comment: Fields used for language detection in the indexer.
     * @return The value of found property. (NotNull: if not found, exception but basically no way)
     */
    String getIndexerLanguageFields();

    /**
     * Get the value for the key 'indexer.language.detect.length'. <br>
     * The value is, e.g. 1000 <br>
     * comment: Length of text for language detection in the indexer.