2717b25931
* Added qr op implementation. Initial version. * Fixed doc for qr op. Signed-off-by: shugeo <sgazeos@gmail.com> * Implementation of QR decomposition. CPU platform version. * Added a pair of tests for qr op testing. Signed-off-by: shugeo <sgazeos@gmail.com> * QR implementation. Signed-off-by: shugeo <sgazeos@gmail.com> * Corrected norm using. * Properly calculated intermediate results with QR decomposition. * Another step to implement QR algorithm by householder. * Cpu implementatio for QR decomposition. The first working edition. * Corrected test to QR decomposition. * Added tad multithreading with QR implementation. * Finished cpu implementation for QR decomposition helpers. * Refactored tests and improved multithreading. * Refactored QR cpu implementation and update cuda implementation helpers. * Cuda QR helper implementation. The first working edition. * Eliminated waste prints. * Restore multithreading with cuda implementation. * Ops names corrected * Refactored qr op helpers to optimize. Signed-off-by: shugeo <sgazeos@gmail.com> * Eliminated waste manual ticking. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored memory allocation to avoid waste memory usage. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored matrixMinor method both for cuda and cpu platforms. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored method of vmul to use raw buffers instead type conversion. Signed-off-by: shugeo <sgazeos@gmail.com> * Refactored temporary array of matricies. Signed-off-by: shugeo <sgazeos@gmail.com> Co-authored-by: Alexander Stoyakin <alexander.stoyakin@gmail.com> Co-authored-by: raver119 <raver119@gmail.com> |
||
---|---|---|
.. | ||
legacy | ||
BarnesHutTsne.cu | ||
README.md | ||
activations.cu | ||
addBias.cu | ||
adjust_hue.cu | ||
adjust_saturation.cu | ||
axis.cu | ||
batched_gemm.cu | ||
batchnorm.cu | ||
betaInc.cu | ||
col2im.cu | ||
compare_elem.cu | ||
concat.cu | ||
confusion.cu | ||
convolutions.cu | ||
cross.cu | ||
d_t_s.cu | ||
diGamma.cu | ||
diag.cu | ||
dilation2d.cu | ||
dropout.cu | ||
dynamic.cu | ||
extract_patches.cu | ||
fake_quantization.cu | ||
flatten.cu | ||
gather.cu | ||
gather_nd.cu | ||
gradient.cu | ||
gru.cu | ||
hamming.cu | ||
hashcode.cu | ||
histogram.cu | ||
histogramFixedWidth.cu | ||
im2col.cu | ||
image_draw_bounding_boxes.cu | ||
image_resize.cu | ||
image_suppression.cu | ||
imagesHelpers.cu | ||
ismax.cu | ||
legacy_helper.cu | ||
lgamma.cu | ||
lrn.cu | ||
lstm.cu | ||
lup.cu | ||
matrixSetDiag.cu | ||
matrix_band.cu | ||
matrix_diag_part.cu | ||
max_pooling.cu | ||
maximum.cu | ||
merge.cu | ||
meshgrid.cu | ||
minimum.cu | ||
nth_element.cu | ||
one_hot.cu | ||
pad.cu | ||
percentile.cu | ||
polyGamma.cu | ||
prefix.cu | ||
print_variable.cu | ||
qr.cu | ||
random.cu | ||
random_crop.cu | ||
range.cu | ||
reverse.cu | ||
roll.cu | ||
s_t_b.cu | ||
s_t_d.cu | ||
scatter.cu | ||
scatter_simple.cu | ||
scatter_update.cu | ||
segment.cu | ||
segment_max.cu | ||
segment_mean.cu | ||
segment_min.cu | ||
segment_prod.cu | ||
segment_sqrtn.cu | ||
segment_sum.cu | ||
sequence_mask.cu | ||
sg_cb.cu | ||
shift.cu | ||
sru.cu | ||
stack.cu | ||
svd.cu | ||
toggle_bits.cu | ||
top_k.cu | ||
transforms.cu | ||
triangular_solve.cu | ||
weights.cu | ||
zeta.cu |
README.md
This folder contains CUDA-specific implementations for operations.