diff options
| author | Sheri Zhang <sheri.zhang@arm.com> | 2021-10-15 19:54:17 +0100 |
|---|---|---|
| committer | Sheri Zhang <sheri.zhang@arm.com> | 2021-10-18 17:36:47 +0000 |
| commit | 5dda2177800009b24e31550ed849b1ef3fca6167 (patch) | |
| tree | dfce69d52db6111d6751f4ee4add6ab172a3290d /docs | |
| parent | c9cecc0e565e7b4978cecc92e03e6c93bb8d0cb9 (diff) | |
| download | ComputeLibrary-5dda2177800009b24e31550ed849b1ef3fca6167.tar.gz | |
DirectConv3d support refine
- Decouple data support of CpuDirectConv3dKernel
- Update documentation for Conv3d
Signed-off-by: Sheri Zhang <sheri.zhang@arm.com>
Change-Id: I1d94aa28f821f45a1a3d39cc3335c8faeee89f0d
Reviewed-on: https://review.mlplatform.org/c/ml/ComputeLibrary/+/6453
Reviewed-by: Giorgio Arena <giorgio.arena@arm.com>
Comments-Addressed: Arm Jenkins <bsgcomp@arm.com>
Tested-by: Arm Jenkins <bsgcomp@arm.com>
Diffstat (limited to 'docs')
| -rw-r--r-- | docs/user_guide/data_layout.dox | 3 | ||||
| -rw-r--r-- | docs/user_guide/operator_list.dox | 3 | ||||
| -rw-r--r-- | docs/user_guide/release_version_and_change_log.dox | 8 |
3 files changed, 12 insertions, 2 deletions
diff --git a/docs/user_guide/data_layout.dox b/docs/user_guide/data_layout.dox index 97d3ea6262..ae69bbf457 100644 --- a/docs/user_guide/data_layout.dox +++ b/docs/user_guide/data_layout.dox @@ -34,8 +34,9 @@ the right-most letter represents the fastest changing dimension: - NHWC: The native layout of Compute Library that delivers the best performance where channels are in the fastest changing dimension - NCHW: Legacy layout where width is in the fastest changing dimension +- NDHWC: New data layout for supporting 3D operators -, where N = batch, C = channel, H = height, W = width. +, where N = batch, C = channel, H = height, W = width, D = depth. */ } // namespace diff --git a/docs/user_guide/operator_list.dox b/docs/user_guide/operator_list.dox index ebc970d8c1..1d06a394a9 100644 --- a/docs/user_guide/operator_list.dox +++ b/docs/user_guide/operator_list.dox @@ -52,9 +52,10 @@ Compute Library supports the following data layouts (fast changing dimension fro <ul> <li>NHWC: The native layout of Compute Library that delivers the best performance where channels are in the fastest changing dimension <li>NCHW: Legacy layout where width is in the fastest changing dimension + <li>NDHWC: New data layout for supporting 3D operators <li>All: Agnostic to any specific data layout </ul> -where N = batches, C = channels, H = height, W = width +where N = batches, C = channels, H = height, W = width, D = depth <table> <caption id="multi_row"></caption> diff --git a/docs/user_guide/release_version_and_change_log.dox b/docs/user_guide/release_version_and_change_log.dox index 583cf4fb82..2470b45203 100644 --- a/docs/user_guide/release_version_and_change_log.dox +++ b/docs/user_guide/release_version_and_change_log.dox @@ -40,6 +40,14 @@ If there is more than one release in a month then an extra sequential number is @section S2_2_changelog Changelog +v21.11 Public major release + - Various bug fixes. + - Various optimizations. + - New OpenCL kernels / functions: + - @ref CLConv3D + - New Arm® Neon™ kernels / functions: + - @ref NEConv3D + v21.08 Public major release - Various bug fixes. - Various optimizations: |