Python bindings for ArrayFire: A general purpose GPU library.
Repo arrayfire/arrayfire-python, read from GitHub 1 min ago. arrayfire has not joined Take My Repo.
fp16 - half precision floating point support has been added - #221
Confidence Connected Components confidenceCC - #221
Deconvolutions - #221
Reduction using keys - #221
Neural network based convolution and gradient functions - #221
Support for uniform ranges in approx1 and approx2 functions - #234
Array class methods - #233
New Examples
create_sparse_from_host - #198bench_cg example on less capable GPUs - #200Array.device_ptr function documentation - #210anisotropic_diffusion(): Anisotropic diffusion filter.topk(): Returns top-K elements given an array.sift() and gloh(), which were improperly calling the library.len() method, which returns array.elements().sign() documentation.helloworld example to match C++ lib.Includes fix for arrayfire.canny.
Feature parity with ArrayFire 3.5.
canny: Canny Edge detectorArray.scalar: Return the first element of the arraydot: Now support option to return scalarprint_mem_info: Prints memory being used / locked by arrayfire memory manager.Array.allocated: Returs the amount of memory allocated for the given buffer.set_fft_plan_cache_size: Sets the size of the fft plan cache.Bug Fixes:
sort_by_key had key and value flipped in documentation.Improvements and bugfixes from upstream include:
approx1.hamming_matcher and nearest_neighbour.