mirrors / dmlc--dgl

项目文件夹

dmlc--dgl
文件
0.6.x
T
文件历史
Quan (Andy) Gan 05c53ca38e [Performance] Prefer parallelized conversion to CSC from COO instead of transposing CSR (#2793) 
* fix coo2csr speed

* add comments
2021-04-05 16:38:36 +08:00
..
api
[Data] Support HeteroGraph save/load (#1526)
2020-06-30 13:34:50 +08:00
array
[Bug] Disable cpu fp16 (#2783)
2021-03-25 18:04:39 +08:00
dataloading
[Dataloading] Add class for copying tensors to/from the GPU on a non-default stream (#2284)
2020-10-30 07:29:06 -07:00
geometry
[Feature] Add edge coarsening for homogeneous undirected graphs (#2691)
2021-03-09 16:35:27 +08:00
graph
[Performance] Prefer parallelized conversion to CSC from COO instead of transposing CSR (#2793)
2021-04-05 16:38:36 +08:00
kernel
[Feature] Add edge coarsening for homogeneous undirected graphs (#2691)
2021-03-09 16:35:27 +08:00
random
[Performance] Linear UniformChoice optimization (#2710)
2021-04-01 16:29:49 +08:00
rpc
Print error messages when using TCP socket (#2763)
2021-03-21 21:00:22 -07:00
runtime
[Bugfix] Wrap cub with CUB_NS_PREFIX and remove dependency on Thrust to linking issues with Torch 1.8 (#2758)
2021-03-22 20:06:57 +08:00
scheduler
[Misc] Move many deprecated codes to the deprecate folder (#1893)
2020-07-30 18:17:33 +08:00
bcast.cc
[Feature] Autograd of gspmm and gsddmm on PyTorch/MXNet/Tensorflow (#1680)
2020-07-13 23:21:22 +08:00
c_api_common.cc
[Hetero] Heterograph C++ implementation; Bipartite and Python wrapper (#725)
2019-07-27 15:22:56 -04:00
c_api_common.h
[CUDA][Kernel] More CUDA kernels; Standardize the behavior for sorted COO/CSR (#1704)
2020-06-28 18:37:28 +08:00