mzl
|
5a5340d03b
remove UtilsBuilder load, use torch (un)flatten ops (#3728)
|
1 年之前 |
Olatunji Ruwase
|
dd8df20fe0
zero3 checkpoint frozen params (#3205)
|
1 年之前 |
Michael Wyatt
|
b361c72761
Update DeepSpeed copyright license to Apache 2.0 (#3111)
|
1 年之前 |
Jeff Rasley
|
91d63e0228
update formatter version and style settings (#3098)
|
1 年之前 |
Olatunji Ruwase
|
541e423ae6
Enable tensor fragments for zero 2 & 3 (#2727)
|
1 年之前 |
Olatunji Ruwase
|
799120e7e4
Universal checkpoint for zero stage 1 (#2284)
|
2 年之前 |
Olatunji Ruwase
|
f4a92a19a6
Checkpoint backwards-compatbility workaround (#2384)
|
2 年之前 |
Olatunji Ruwase
|
53182531ed
Refactor universal checkpointing and tensor fragments (#2253)
|
2 年之前 |
Mikhail Druzhinin
|
4671cce558
Fix OrderedDict import for python3.6 (#2267)
|
2 年之前 |
shjwudp
|
57140e8e95
fix: fix BF16_Optimizer compatibility issue with optimizer state 0-dim tensor (#2152)
|
2 年之前 |
Alex Hedges
|
316c4a43e0
Add flake8 to pre-commit checks (#2051)
|
2 年之前 |
Olatunji Ruwase
|
80d0a32f0b
Checkpoint reshaping (#1953)
|
2 年之前 |
Karim Foda
|
735406e536
fix import errors (#2026)
|
2 年之前 |
Ammar Ahmad Awan
|
36ad3119d5
DeepSpeed comm backend v1 (#1985)
|
2 年之前 |
Jeff Rasley
|
50893458d6
Fairseq support (#1915)
|
2 年之前 |
Olatunji Ruwase
|
af58f63dde
bf16 inference (#1917)
|
2 年之前 |
Olatunji Ruwase
|
56c5223868
bf16+pipeline parallelism (#1801)
|
2 年之前 |