train

void train(Index::idx_t n, const float *x) override

Trains the quantizer of AscendIndexSQ. It inherits the APIs in AscendFaiss and provides specific implementation. You must train the index before using the add method.

Index::idx_t n: number of feature vectors in the training dataset.

const float *x: feature vector data.

N/A

N/A

• The value of n cannot be greater than 1e9.
• The x must be a non-null pointer with a length of dim x n. Otherwise, an out-of-bounds read/write error may occur, causing program breakdown.
• Data distribution is collected during training. If the training dataset is small, the query accuracy may be affected.