此网站需要 JavaScript。
96cd2ee653
[Performance][GPU] Improve _SegmentCopyKernel() (#3470 )
2021-11-06 10:53:53 -07:00
efe0b06162
fix test dataloader (#3482 )
2021-11-05 15:13:56 +08:00
582f71a173
Fix openmp header (#3325 )
2021-09-13 15:41:28 +08:00
c5ae54bfe6
[Doc] Evaluation Tutorial for Link Prediction (#3463 )
2021-11-05 13:09:56 +08:00
0d1807ecf4
update version
2021-11-05 03:12:27 +00:00
50dd8517c8
Enable faster validation for pytorch graphsage example (#3361 )
2021-09-19 17:25:46 -07:00
46557ee8a8
Update cub for cuda 11.5 compatibility (#3468 )
2021-11-02 22:14:18 -07:00
b717c8bf0d
[BugFix] Fix bugs in GPU sampling and enable unit tests for dataloaders on the GPU (#3474 )
2021-11-05 01:34:56 +08:00
d3ae7544cd
[Feature] aten::Relabel_() for the GPU (#3445 )
2021-11-04 23:58:31 +08:00
ef1b7175d5
fix compatibility with PyTorch 1.10 (#3454 )
2021-10-29 09:33:08 +08:00
87a4cb363a
Fix #3437 (#3440 )
2021-10-26 17:56:48 +08:00
ad19562a30
[Doc] remove duplicate papers (#3393 )
2021-10-19 01:41:37 -05:00
e3dd242072
[Fix] Split nccl sparse push into two groups (#3404 )
2021-10-18 03:54:09 -07:00
c7f94dbf5a
[BugFix] Avoid Memory Leak Issue in PyTorch Backend (#3386 )
2021-10-18 04:24:54 -05:00
e87dec8599
[Bugfix][Pytorch] Fix model save and load bug of stgcn_wave (#3303 )
2021-10-18 10:17:54 +02:00
69eae6de49
[Fix] Use ==/!= to compare constant literals (str, bytes, int, float, tuple) (#3415 )
2021-10-14 11:35:37 +02:00
a63ad14bda
[PyTorch][Bugfix] Use uint8 instead of bool in pytorch to be compatible with nightly version (#3406 )
2021-10-14 02:34:55 -07:00
71220a12db
[Bugfix] three bugs related to using DGL as a subdirectory(third_party) of another project. (#3379 )
2021-10-14 14:43:23 +08:00
bf11b19e76
Update dataloader.py (#3322 )
2021-09-06 12:16:03 +08:00
f46080a4d1
[Feature] k-hop Subgraph Extraction (#3458 )
2021-11-04 15:47:35 +08:00
23d575ba81
[Bug] Do not skip graphconv even no edge exists (#3416 )
2021-10-14 14:34:42 +08:00
ee183f69a1
[BugFix] add count_nonzero() into SA_Client (#3417 )
2021-10-12 21:17:30 +08:00
cc83b49e01
[Bug] check dtype before convert to gk (#3414 )
2021-10-12 14:34:40 +08:00
67e3902711
[Model] Refine GraphSAINT (#3328 )
2021-10-07 19:06:28 +08:00
56b04b2125
[BugFix] extract gz into target dir (#3389 )
2021-09-30 11:35:40 +08:00
ebd2c08f99
[Distributed] Allow user to pass-in extra env parameters when launching a distributed training task. (#3375 )
2021-09-23 19:03:33 +08:00
3d90edda3e
Fix torch import in example (#3372 )
2021-09-22 22:23:05 -07:00
84169b1954
[Feature] Graceful handling of exceptions thrown within OpenMP blocks (#3353 )
2021-09-22 13:44:46 +08:00
039856f8a9
[Doc] Added md5sum info for OGB-LSC dataset (#3332 )
2021-09-21 09:36:53 -05:00
6df6a0881e
[BugFix] fix typo in fakenews dataset variable name (#3363 )
2021-09-21 09:06:26 -05:00
9b597b8e74
fix broadcast tensor dim (#3351 )
2021-09-19 06:49:56 -07:00
1c9274f168
[BugFix] initialize data if null when converting from row sorted coo to csr (#3360 )
2021-09-17 17:07:05 +08:00
2b45b8c72d
[Performance] improve coo2csr space complexity when row is not sorted (#3326 )
2021-09-14 10:41:55 +08:00
50082a571f
[Performance, CPU] Rewriting OpenMP pragmas into parallel_for (#3171 )
2021-09-02 10:39:48 +02:00
f9d51fdf0b
[Feature] Add a HINT for the per edge type sampler of heterogeneous DistGraph that highlighting the etypes are sorted already. (#3260 )
2021-09-01 15:36:01 +08:00
b0346bbd4f
[CPU][Sampling][Performance] Improve sampling on the CPU. (#3274 )
2021-08-31 13:37:49 -07:00
ba67fa0fc2
[Bugfix] And PYTHONPATH in server launch. (#3352 )
2021-09-14 17:59:16 +08:00
269f90428c
[Bugfix] Fix #3291 (#3333 )
2021-09-14 08:57:20 +08:00
a171f41270
Fixes bug #3312 (#3345 )
2021-09-13 20:31:35 +05:30
f77bee328a
[Bugfix] Fix bugs of farthest_point_sampler (#3327 )
2021-09-10 17:50:06 +08:00
e3eb3ee511
Fix distributed device mapping problem. (#3313 )
2021-09-02 13:52:00 +08:00
64f20eeaec
Update CONTRIBUTORS.md (#3475 )
2021-11-03 17:05:15 +02:00
e2f33fd5cc
[NN][Model] GATv2 (#3473 )
2021-11-03 15:55:56 +02:00
f5102145a7
Update cub for cuda 11.5 compatibility (#3468 )
2021-11-02 22:14:18 -07:00
95b1eb734f
[Distributed] Fix a bug in sampling an empty frontier (#3298 )
2021-08-29 10:54:46 -07:00
6b6174a5c4
Cleanup some code in graph_services.py (#3238 )
2021-08-16 02:58:37 +08:00
4010e20cd9
[Bugfix] Distributed training can not work with dgl.dataloading.negative_sampler (#3215 )
2021-08-06 20:41:20 +08:00
0246bf5567
[Bugfix] Fix bug introduced by https://github.com/dmlc/dgl/pull/3131 (#3234 )
2021-08-10 11:36:57 +08:00
f0903275e8
[Distributed] Enable distributed EdgeDataLoader (#3192 )
2021-07-28 15:28:27 +08:00
508197e807
[New Feature] Per edge type sampler for to_homogeneous graphs. (#3131 )
2021-07-28 11:51:32 +08:00
223795eb82
[Distributed] Distributed heterograph training (#3069 )
2021-07-17 10:57:00 +08:00
cac25f636f
fix compatibility with PyTorch 1.10 (#3454 )
2021-10-29 09:33:08 +08:00
9067565a25
update contributors (#3451 )
2021-10-28 13:19:21 +08:00
51c6509704
[NN] Add EGATConv nn.module (#3425 )
2021-10-27 18:38:08 +02:00
a9c83bce15
Fix #3437 (#3440 )
2021-10-26 17:56:48 +08:00
579cd3eb49
Update README.md (#3442 )
2021-10-26 02:55:09 -04:00
a8c81018c5
[Sampling] Implement dgl.compact_graphs() for the GPU (#3423 )
2021-10-21 13:07:35 +08:00
308e52a37e
[Doc] remove duplicate papers (#3393 )
2021-10-19 01:41:37 -05:00
c560040f5f
[Fix] Split nccl sparse push into two groups (#3404 )
2021-10-18 03:54:09 -07:00
aa11aaa4ba
[Peformance] Parallelize CSRSliceRows() (#3409 )
2021-10-18 04:25:21 -05:00
ff94ee80b1
[BugFix] Avoid Memory Leak Issue in PyTorch Backend (#3386 )
2021-10-18 04:24:54 -05:00
f703941840
[Bugfix][Pytorch] Fix model save and load bug of stgcn_wave (#3303 )
2021-10-18 10:17:54 +02:00
b81bb91465
[Bug] Fix edge exclusion still not working for full neighbor sampling (#3424 )
2021-10-15 17:51:56 +08:00
4f5c3aa234
[Bugfix] Add UVM specialized IndexSelect kernels which perform boundary checks (#3293 )
2021-10-15 02:05:00 -05:00
04ed6126b5
[Fix] Use ==/!= to compare constant literals (str, bytes, int, float, tuple) (#3415 )
2021-10-14 11:35:37 +02:00
b81efb2b2d
[PyTorch][Bugfix] Use uint8 instead of bool in pytorch to be compatible with nightly version (#3406 )
2021-10-14 02:34:55 -07:00
a47ab71d44
[DOCS] Add training on CPU sections to docs (#3398 )
2021-10-14 08:53:16 +02:00
188630698e
[Bugfix] three bugs related to using DGL as a subdirectory(third_party) of another project. (#3379 )
2021-10-14 14:43:23 +08:00
5d4f6bca2a
[Fix] Fix edge ID exclusion not working in EdgeDataLoader (#3412 )
2021-10-14 14:37:54 +08:00
8798872f54
[Bug] Do not skip graphconv even no edge exists (#3416 )
2021-10-14 14:34:42 +08:00
7c7b60be18
[BugFix] add count_nonzero() into SA_Client (#3417 )
2021-10-12 21:17:30 +08:00
2d88db5a3c
[Bug] check dtype before convert to gk (#3414 )
2021-10-12 14:34:40 +08:00
d947287367
[README] Add GNNLens (#3411 )
2021-10-11 16:27:47 +08:00
532eaa879b
backward now stores DGLGraph index,not DGLGraph object witattached data (#3410 )
2021-10-11 02:08:01 -04:00
aef96dfa34
[Model] Refine GraphSAINT (#3328 )
2021-10-07 19:06:28 +08:00
f9fd7fd7f7
[BugFix] extract gz into target dir (#3389 )
2021-09-30 11:35:40 +08:00
e234fcfa8f
[Feature] enable create/set/free cuda stream for internal use (#3334 )
2021-09-29 15:35:02 +08:00
5cf48fc69c
[Feature] Implement one thread multiple socket (#3200 )
2021-09-28 12:45:52 +08:00
179d6aab1e
[Distributed] Allow user to pass-in extra env parameters when launching a distributed training task. (#3375 )
2021-09-23 19:03:33 +08:00
367a3a34c4
Fix torch import in example (#3372 )
2021-09-22 22:23:05 -07:00
a04a8d066e
[Feature] Graceful handling of exceptions thrown within OpenMP blocks (#3353 )
2021-09-22 13:44:46 +08:00
bc14829fb3
[Feature] Exclude edges in sample_neighbors (#2971 )
2021-09-21 20:15:56 +02:00
ac9261b2a0
[Doc] Added md5sum info for OGB-LSC dataset (#3332 )
2021-09-21 09:36:53 -05:00
fecd6f3c50
[BugFix] fix typo in fakenews dataset variable name (#3363 )
2021-09-21 09:06:26 -05:00
01a2214430
Enable faster validation for pytorch graphsage example (#3361 )
2021-09-19 17:25:46 -07:00
bc5cac4426
fix broadcast tensor dim (#3351 )
2021-09-19 06:49:56 -07:00
bacc9047f2
[BugFix] initialize data if null when converting from row sorted coo to csr (#3360 )
2021-09-17 17:07:05 +08:00
2647afc9b3
[Performance][Feature] Add src_nodes paramter to to_block() to avoid cost running unique() when available. (#2973 )
2021-09-15 18:05:16 -07:00
ad61a9a550
[Bugfix] And PYTHONPATH in server launch. (#3352 )
2021-09-14 17:59:16 +08:00
f4c79f7f6d
[Performance] improve coo2csr space complexity when row is not sorted (#3326 )
2021-09-14 10:41:55 +08:00
a609b4f023
[Bugfix] Fix #3291 (#3333 )
2021-09-14 08:57:20 +08:00
983a4fdd19
Fixes bug #3312 (#3345 )
2021-09-13 20:31:35 +05:30
3fef5d27d3
[Model] PCT (#3339 )
2021-09-13 17:17:24 +08:00
e7ea0f5364
Fix openmp header (#3325 )
2021-09-13 15:41:28 +08:00
26b631805f
[Bugfix] Fix Correct&Smooth (#3329 )
2021-09-13 11:46:14 +08:00
246df1a8f4
update contributors (#3347 )
2021-09-12 00:40:05 +08:00
06f62b99a0
Update contributors (#3346 )
2021-09-11 00:31:18 -07:00
6454c791b8
[Bugfix] Fix bugs of farthest_point_sampler (#3327 )
2021-09-10 17:50:06 +08:00
d17b299a13
[Feature] Update edge features (apply_edge) for heterogeneous graph (#3250 )
2021-09-09 10:39:08 -04:00
129f67cd91
[Jenkins] add Rhett-Ying into authorized trigger (#3331 )
2021-09-08 14:53:58 +08:00
ayasar70