Skip to content
GitLab
Explore
Sign in
Register
wanggh
apex
Repository
Branches
Overview
Active
Stale
All
radix_decomposition
33108ae4
·
Radix decomposition cuda kernel
·
Dec 31, 2019
apex_for_20.01_pip_war
c2fe75d8
·
Fix pip (#655)
·
Dec 15, 2019
batchnorm_1d_patch
acea19df
·
fixing batchnorm 1d input
·
Nov 06, 2019
gh-pages
2712abca
·
Generated gh-pages for commit
08898593
·
Oct 10, 2019
remove_nvcc
956991d3
·
removing nvtx range used for debugging
·
Sep 10, 2019
data_ptr_fix
1e522ad7
·
Removing deprecated checks
·
Sep 05, 2019
multi_tensor_sgd
f2b16db0
·
Merge pull request #445 from ptrblck/multi_tensor_sgd
·
Aug 21, 2019
apex-ci
6be1c616
·
Add directory
·
Jul 25, 2019
fix_non_affine_fused_layer_norm
c32d822c
·
Bug fix for non-affine layer-norm + add backward unit test
·
Jul 25, 2019
persistent_sync_bn_group8_fix
89ae9e54
·
Fixing rank mapping for bn_group size == 8
·
Jul 12, 2019
sbn_fix_PR
23026b35
·
[sbn update]
·
Jun 28, 2019
lamb_add_fp16_support_update_term
3aeea0d8
·
Add support for fp16 update term (new UPD_T typename in template)
·
Jun 28, 2019
gbn_update
333e53f7
·
commenting on unnecessarily exposed buffer at user code
·
Jun 17, 2019
prioritize_custom_to
af682b4c
·
Give custom to method a higher priority
·
Jun 17, 2019
slightly_faster_lamb_kernels
5d5f7b86
·
Separate LDG/STG from compute loop
·
Jun 13, 2019
option_disable_allreduce_in_DDP
373f4461
·
Add option to turn on/off allreduce in DDP (useful for gradient accumulation)
·
Jun 12, 2019
l2norm_for_bert
4951442c
·
ILP for l2norm functor
·
May 30, 2019
multi_tensor_lamb_optimizer
ef005453
·
Fix compilation errors
·
May 30, 2019
bnp_test
1f45735d
·
added unit test for BNP
·
May 21, 2019
add_param_group
4d32cb8c
·
Test added and passing
·
May 15, 2019
Prev
1
2
3
4
5
6
Next