This API implements
$$y = \frac{x}{||x||_p}$$
This API divides each element in x by p-norm and assigns the results to elements in y.
API signatures:
gscaleNorm( N, p, x, strideX, y, strideY )
gscaleNorm.ndarray( N, p, x, strideX, offsetX, y, strideY, offsetY )
where
N: number of indexed elements.p: order of the norm.x: input strided array.strideX: stride length for x.offsetX: starting index for x.y: output strided array.strideY: stride length for y.offsetY: starting index for y.
Notes
- To implement, if provided a special value of
p, delegate to specialized APIs; otherwise, compute the p-norm, followed by blas/ext/base/*wax with alpha = 1 / norm(x).
p should be greater than 0, with the single exception being -infinity. For this case, normalize by minabs(x).
Follow-on Work
Once completed and merged, the following packages can be added
This API implements
This API divides each element in
xby p-norm and assigns the results to elements iny.API signatures:
where
N: number of indexed elements.p: order of the norm.x: input strided array.strideX: stride length forx.offsetX: starting index forx.y: output strided array.strideY: stride length fory.offsetY: starting index fory.Notes
p, delegate to specialized APIs; otherwise, compute the p-norm, followed byblas/ext/base/*waxwithalpha = 1 / norm(x).pshould be greater than0, with the single exception being-infinity. For this case, normalize byminabs(x).Follow-on Work
Once completed and merged, the following packages can be added
blas/ext/base/dscale-normblas/ext/base/sscale-normblas/ext/base/ndarray/dscale-normblas/ext/base/ndarray/sscale-normblas/ext/base/ndarray/gscale-normblas/ext/scale-norm