rpp
rpp copied to clipboard
AMD ROCm Performance Primitives (RPP) library is a comprehensive high-performance computer vision library for AMD processors with HIP/OpenCL/CPU back-ends.
Fixes broken links / images (#331) in 6.1.0 docs. ------- * added missing outputs for image augmentations modified the description with correct output names * added gif for voxel input...
Fixes broken links/images (#331) in 6.1.1 docs #346 ------- * added missing outputs for image augmentations modified the description with correct output names * added gif for voxel input and...
### Problem Description Does RPP require the graphics use case or just the rocm use case? If this does require graphics use case what are the functionality support or rocm...
This is a minor PR to add some HIP fixes to match-up tensor code for uniformity: - Uses pre-allocated common scratchBufferHip everywhere in Tensor code for scratch HIP memory -...
@rrawther @kiritigowda MelfilterBank as a functionality is dependent on spectrogram. So PR #312 needs to be merged into develop first, followed by a develop/ToT sync to ensure only MelFilterBank files-changed...
- Adds Tensor Audio support on HOST for Spectrogram - Adds relevant QA pass/fail tests and Performance tests
This PR modifies the already added voxel slice kernels to now accept anchor and shape params for a more generic version for Audio support. Slice can now function in all...
- Adds support for Tensor Normalize ND on HOST and HIP - Adds relevant Unit/Perf/QA tests
- Adds Erase support for HOST and HIP - Adds relevant unit/perf/QA tests
- Adds support for remap on an NHWC or NCHW layout tensor on HOST and HIP - Adds support for U8/F16/F32/I8 bit depths and layout toggle variants - Adds support...