Commits

Commits on May 14, 2024

[manifest] update mainfest to add hpp file in deepspeed. (#5533)
yisustc
authored
ebf82e8
Switch from double quotes to match single quotes (#5530)
loadams
authored
62ca317

Commits on May 13, 2024

Switch pynvml to nvidia-ml-py (#5529)
loadams
authored
82ce4ae
enable phi2 autotp (#5436)

authored
3a7f3aa
[manifest] update mainfest to add hpp file in csrc. (#5522)
yisustc
authored
4696afd

Commits on May 10, 2024

Fused adam for HPU (#5500)

BacharL
and
tjruwase
authored
df4ef0a

Commits on May 8, 2024

enable phi3_mini autotp (#5501)

Yejing-Lai
and
lekurile
authored
3dd7ccf
Fix compile wrapper (#5455)

authored
0b224ed

Commits on May 7, 2024

Fix crash when creating Torch tensor on NPU with device=get_accelerator().current_device() (#5464)

authored
0fc19b6

Commits on May 3, 2024

re-introduce: stage3: efficient compute of scaled_global_grad_norm (#5493)
nelyahu
authored
90793aa

Commits on Apr 29, 2024

Commits on Apr 24, 2024

Add getter and setter methods for compile_backend across accelerators. (#5299)

authored
fa8458b

Commits on Apr 23, 2024

Update version.txt after 0.14.2 release (#5458)

mrwyattii
and
loadams
authored
fbdf0ea
Update PyTest torch version to match PyTorch latest official (2.3.0) (#5454)
loadams
authored
5f631ab
Update README.md (#5453)

Jhonso7393
and
loadams
authored
ad20279
OptimizedLinear implementation (#5355)

authored
5e6c9b9

Commits on Apr 22, 2024

set the default to use set_to_none for clearing gradients in BF16 optimizer. (#5434)

inkcherry
and
loadams
authored
c66bc42
Improve parallel process of universal checkpoint conversion (#5343)

tohtana
and
tjruwase
authored
c292b03
64bit indexing fused adam (#5187)

authored
9b6ef9e

Commits on Apr 20, 2024

add device config env for the accelerator (#5396)

authored
3f875d9

Commits on Apr 19, 2024

Fix sorting of shard optimizer states files for universal checkpoint (#5395)

tohtana
and
tjruwase
authored
99951ca
Selective dequantization (#5375)

authored
c632ea0

Commits on Apr 18, 2024

Parallel map step for `DistributedDataAnalyzer` map-reduce (#5291)

authored
64defe6
Bug fix in `split_index` method (#5292)

bm-synth
and
loadams
authored
aaaf8bc
Add required paths to trigger AMD tests on PRs (#5406)

loadams
and
root
authored
3194fe8

Commits on Apr 17, 2024

[CPU] Support SHM based inference_all_reduce in TorchBackend (#5391)

authored
b22706a
Add more tests into XPU CI (#5427)
Liangliang-Ma
authored
4d4c99d

Commits on Apr 16, 2024

Update _sidebar.scss (#5293)

fasterinnerlooper
and
loadams
authored
a9cbd68
Update engine.py to avoid torch warning (#5408)

authored
bc0f774
Update 'create-pr' action in release workflow to latest (#5415)
loadams
authored
34969d2