VEDA Tensors Library

VEDA Tensors is a library to provide generic usable high performance kernels usually used in AI frameworks. VEDA Tensors is still beta and far from being complete.

Release Notes

Version	Comment
v0.1.7	Fixed race condition in cat
v0.1.6	Added VE3 support
v0.1.5	Added unary Not Bugfixes for NCC >= 5
v0.1.4	Upgraded to VEDA 2.0 Added Log1P
v0.1.3	Fixed dtype checking in veda_tensors_unary_c
v0.1.2	Added L0, L1, L2 to veda_tensors_reduce_dim Added veda_tensors_unary_ts Added veda_tensors_unary_tss Added veda_tensors_unary_ttt Added veda_tensors_where Added veda_tensors_binary_s Added veda_tensors_softmax Added VEDA_TENSORS_UNARY_CLAMP Added VEDA_TENSORS_UNARY_POW Improved performance of several functions.
v0.1.1	Added AdaDelta, AdaGrad, Adam and AdaMax optimizer functions
v0.1.0	Initial Release