Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
76 commits
Select commit Hold shift + click to select a range
230f479
[ROCm] Enable fp_quantizer on ROCm (#7027)
rraminen Feb 13, 2025
130c11c
Add pyproject.toml
loadams Feb 13, 2025
b2e16cb
Formatting fix
loadams Feb 13, 2025
72bfb70
add gds chinese blog (#7034)
GuanhuaWang Feb 13, 2025
48c02ad
Add chinese blog for deepspeed windows, and fix format (#7035)
hwchen2017 Feb 14, 2025
4996cae
AIO on ROCM (#7023)
jomayeri Feb 14, 2025
52709fa
Update setuptools min requirement
loadams Feb 14, 2025
9eb0618
Switch build to legacy
loadams Feb 14, 2025
36ce373
Add no-build isolation
loadams Feb 14, 2025
a0ff11a
Control trace cache warnings (#7039)
tjruwase Feb 18, 2025
c574b45
Update CUDA compute capability to support Blackwell (#7047)
hwchen2017 Feb 18, 2025
5326873
Update setup.py handling of ROCm cupy (#7051)
loadams Feb 19, 2025
c22be1a
nv-ds-chat breaks with latest transformers (#7052)
loadams Feb 19, 2025
2873a11
Test with non legacy backend
loadams Feb 19, 2025
4e88463
Need legacy backend to execute setup.py, though it executes it differ…
loadams Feb 19, 2025
6d837ca
Update to actually use legacy backend
loadams Feb 19, 2025
5b2f713
Rename aio_thread_count to intra_op_parallelism (#7056)
tjruwase Feb 19, 2025
c0f4235
add autoTP training zero2 tests (#7049)
inkcherry Feb 19, 2025
01dd318
Fix, bf16 optimizer remove dup loop (#7054)
wukong1992 Feb 20, 2025
41718ad
Update version.txt after 0.16.4 release (#7063)
loadams Feb 20, 2025
c83ade6
fix an outdated doc wrt CUDA_VISIBLE_DEVICES (#7058)
stas00 Feb 20, 2025
cc79d17
Tecorigin sdaa accelerator (#6903)
siqi654321 Feb 20, 2025
aa63ba4
Handle special case of libuv for Windows (#7064)
loadams Feb 20, 2025
aaf86d6
Update README with info on newest accelerator (#7065)
loadams Feb 21, 2025
630864d
Bug Fix for offload_states API (#7050)
U-rara Feb 21, 2025
5a2ecea
Add all imports from setup.py to pyproject.toml
loadams Feb 21, 2025
ac52bae
Fix TOCTOU issues, switch to fstat (#7067)
loadams Feb 24, 2025
fd5273f
config torch to avoid graph breaks caused by logger (#6999)
ShellyNR Feb 24, 2025
c91d225
Fix meta load tensor imcompatible issue (#7073)
Yejing-Lai Feb 24, 2025
80ff420
Replace calls to `python setup.py sdist` with `python -m build --sdis…
loadams Feb 24, 2025
355e9c5
Update workflow to not do build isolation
loadams Feb 25, 2025
00addaa
Add more requirements
loadams Feb 25, 2025
432e71d
Revert "Handle special case of libuv for Windows (#7064)" (#7076)
loadams Feb 25, 2025
5ab8589
Add DeepseekV3 AutoTP. (#7045)
Yejing-Lai Feb 26, 2025
f325c92
Add pyproject.toml into comments
loadams Feb 26, 2025
9c4c444
Improve inference tutorial docs (#7083)
loadams Feb 26, 2025
5ebe9bd
Pin transformers version on tests that use latest. (#7085)
loadams Feb 27, 2025
9da1b33
Update README.md with ICS '23 MoE paper link (#7087)
siddharth9820 Feb 27, 2025
f963248
Update parallelism for nv-torch-latest/nightly tests due to more GPUs…
loadams Feb 27, 2025
8254e1b
Remove workflows for very old torch versions (#7090)
loadams Feb 28, 2025
a7ab4ab
Use new dlpack api; Formatting fixes (#7101)
tjruwase Mar 3, 2025
6eb5baa
Avoid graph breaks by disabling sourceless calls in instrument_w_nvtx…
deepcharm Mar 3, 2025
feeaa2e
Avoid graph breaks in torch.compile caused by inner classes in the ba…
deepcharm Mar 4, 2025
7b7e112
Only run pre-commit on the changes (#7106)
hwchen2017 Mar 4, 2025
674f9bd
Avoid graph break due to unsupported frozenset (#7105)
deepcharm Mar 4, 2025
240c2a7
Fix fused_qkv print model ValueError (#7109)
Yejing-Lai Mar 4, 2025
1c35241
Update references to new X/Twitter handle (#7110)
loadams Mar 4, 2025
ec2e93e
Update gaudi2 nightly,ci to latest 1.20.0 build (#7093)
raza-sikander Mar 7, 2025
c793401
fix keep_module_on_host (#7112)
inkcherry Mar 10, 2025
62abd88
Add sequential pytest mark to TestNVMeCheckpointing to resolve pytest…
loadams Mar 11, 2025
065084b
Training multiple models (#7018)
tjruwase Mar 11, 2025
e93f0d1
Update CONTRIBUTING.md to reflect changes from CLA to DCO (#7135)
loadams Mar 14, 2025
eaeae79
Avoid missing attr error (#7133)
tjruwase Mar 14, 2025
8df6245
Add conditional expression (#7119)
A-transformer Mar 14, 2025
6de20f6
Unpin transformers version for most workflows (#7139)
loadams Mar 14, 2025
e0188ac
Conditionally quote env vars (#7071)
saurabhkoshatwar Mar 17, 2025
ea61c6c
Correct the BACKWARD_PREFETCH_SUBMIT mismatch (#7120)
A-transformer Mar 17, 2025
9d75311
Enhance Gaudi2 CI/Nightly Coverage with Model Parallelism and Linear …
raza-sikander Mar 18, 2025
4ca7ba5
Update container version that runs on A6000 tests. (#7153)
loadams Mar 19, 2025
2db922f
fix leak of z3 buffer
Mar 20, 2025
78ec025
hf tp+zero training doc. (#7151)
inkcherry Mar 20, 2025
4a851f3
Avoid graph break by removing redundant requires_grad attr change (#7…
deepcharm Mar 24, 2025
6adcf2e
Add destroy to tests to free memory (#7160)
tohtana Mar 24, 2025
7b7ac9e
[NFC] Typo fix in SP layer. (#7152)
c8ef Mar 24, 2025
31ec2b7
Link AutoTP blog in the front page (#7167)
hwchen2017 Mar 25, 2025
e40df22
Merge branch 'master' into loadams/pyproject-toml
loadams Mar 25, 2025
4c32a9d
Remove unneeded requires in build system declaration
loadams Mar 25, 2025
c4b24fd
Add build to the pyproject
loadams Mar 25, 2025
b330b4e
Add no torch build triggers
loadams Mar 28, 2025
3601c29
Remove no build isolation from nv-torch-latest
loadams Mar 28, 2025
e0d9ba4
Merge branch 'master' into loadams/pyproject-toml
loadams Mar 28, 2025
42e42a0
Merge branch 'master' into loadams/pyproject-toml
loadams Apr 3, 2025
3c94f51
Merge branch 'master' into loadams/pyproject-toml
loadams Apr 9, 2025
221f2be
Merge branch 'master' into loadams/pyproject-toml
loadams Apr 9, 2025
dfcc24d
Merge branch 'master' into loadams/pyproject-toml
loadams Apr 18, 2025
b4ed94a
Merge branch 'master' into loadams/pyproject-toml
loadams Apr 21, 2025
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions .github/workflows/no-torch.yml
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,8 @@ on:
- 'accelerator/**'
- '.github/workflows/no-torch.yml'
- 'op_builder/**'
- 'setup.py'
- 'pyproject.toml'
schedule:
- cron: "0 0 * * *"

Expand Down
1 change: 1 addition & 0 deletions .github/workflows/nv-mii.yml
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,7 @@ on:
- '.github/workflows/nv-mii.yml'
- 'requirements/**'
- 'setup.py'
- 'pyproject.toml'
- 'deepspeed/__init__.py'
- 'deepspeed/inference/**'
- '!deepspeed/inference/v2/**' # exclude v2 dir
Expand Down
2 changes: 1 addition & 1 deletion .github/workflows/nv-pre-compile-ops.yml
Original file line number Diff line number Diff line change
Expand Up @@ -36,7 +36,7 @@ jobs:
#python -c "import torch; print('CUDA available:', torch.cuda.is_available())"
- name: Compile DeepSpeed Ops
run: |
DS_ACCELERATOR=cuda DS_ENABLE_NINJA=1 TORCH_CUDA_ARCH_LIST="7.0;7.5;8.0" DS_BUILD_OPS=1 DS_BUILD_SPARSE_ATTN=0 DS_BUILD_FP_QUANTIZER=0 DS_BUILD_CUTLASS_OPS=0 DS_BUILD_GDS=0 DS_BUILD_RAGGED_DEVICE_OPS=0 DS_BUILD_EVOFORMER_ATTN=0 DS_BUILD_DEEP_COMPILE=0 pip3 install .
DS_ACCELERATOR=cuda DS_ENABLE_NINJA=1 TORCH_CUDA_ARCH_LIST="7.0;7.5;8.0" DS_BUILD_OPS=1 DS_BUILD_SPARSE_ATTN=0 DS_BUILD_FP_QUANTIZER=0 DS_BUILD_CUTLASS_OPS=0 DS_BUILD_GDS=0 DS_BUILD_RAGGED_DEVICE_OPS=0 DS_BUILD_EVOFORMER_ATTN=0 DS_BUILD_DEEP_COMPILE=0 pip3 install . --no-build-isolation
- name: DS Report
run: |
ds_report
2 changes: 1 addition & 1 deletion .github/workflows/nv-torch-nightly-v100.yml
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,7 @@ jobs:

- name: Install deepspeed
run: |
pip install .[dev,1bit,autotuning]
pip install .[dev,1bit,autotuning] --no-build-isolation
ds_report

- name: Python environment
Expand Down
2 changes: 1 addition & 1 deletion deepspeed/git_version_info.py
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@
# DeepSpeed Team

try:
# This is populated by setup.py
# This is populated by setup.py/pyproject.toml
from .git_version_info_installed import * # noqa: F401 # type: ignore
except ModuleNotFoundError:
import os
Expand Down
9 changes: 9 additions & 0 deletions pyproject.toml
Original file line number Diff line number Diff line change
@@ -0,0 +1,9 @@
[build-system]
requires = [
"build",
"setuptools>=64",
"torch",
"wheel"
Comment on lines +4 to +6

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If you depend on setuptools 70.1 or later, you won't need wheel.

Suggested change
"setuptools>=64",
"torch",
"wheel"
"setuptools>=70.1",
"torch"

]
# Use legacy backend to import local packages in setup.py
build-backend = "setuptools.build_meta:__legacy__"
Loading