[zero] revert PR #3166, it disabled grad clip for bf16 #3790
zero++ tutorial PR (#3783)
df1859d6
[Fix] _conv_flops_compute when padding is a str and stride=1 (#3169)
d81a6ad6
fix interpolate flops compute (#3782)
a8c182a4
use `Flops Profiler` to test `model.generate()` (#2515)
c4c442f0
revert PR #3166, it disabled grad clip for bf16
9bd7b24c
ensure no loss scaling for non-fp16 dtypes
6075a29d
revert PR #3611 (#3786)
fc9e1ee0
bump to 0.9.6
40045dc7
Merge branch 'master' into revert-3166
710a59c6
ZeRO++ chinese blog (#3793)
49a0a1bb
remove staging trigger (#3792)
2c62cb4c
DeepSpeed-Triton for Inference (#3748)
4dc65f7b
ZeRO++ (#3784)
e1119d8b
adding zero++ to navigation panel of deepspeed.ai (#3796)
01b843aa
Add ZeRO++ Japanese blog (#3797)
319b64ed
Bug Fixes for autotuner and flops profiler (#1880)
b4a2c0af
Missing strided copy for gated MLP (#3788)
b7e1010b
Requires grad checking. (#3789)
e5b1eadb
bump to 0.10.0
9c756cf8
Fix Bug in transform.cu (#3534)
a204edc7
bug fix: triton importing error (#3799)
f6e2e38b
Merge branch 'master' into revert-3166
5c8bae02
jeffra
force-pushed the
master
branch
from
f6e2e38b
to
bafaf3c0
2 years ago
Merge branch 'master' into revert-3166
928dc2c6
tjruwase
approved these changes
on 2023-06-26
Merge branch 'master' into revert-3166
c290d4c1
loadams
approved these changes
on 2023-06-26
Merge branch 'master' into revert-3166
25e083a1
Merge branch 'master' into revert-3166
cafd8186
Merge branch 'master' into revert-3166
f3c44ccc
Merge branch 'master' into revert-3166
4854b5c0
Merge branch 'master' into revert-3166
a8ffc37f
tjruwase
merged
691d246e
into master 2 years ago
mrwyattii
deleted the revert-3166 branch 2 years ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub