.. |
autotuning
|
a32789b033
Modified regular expression (#5306)
|
6 months ago |
checkpoint
|
c292b03a40
Improve parallel process of universal checkpoint conversion (#5343)
|
6 months ago |
comm
|
b22706a721
[CPU] Support SHM based inference_all_reduce in TorchBackend (#5391)
|
6 months ago |
compression
|
389bf69319
fix: Remove duplicate word the (#4051)
|
1 year ago |
elasticity
|
6dcced1d5c
Cleanup required_torch_version code and references. (#5370)
|
6 months ago |
inference
|
e3d873a00e
Fix the FP6 kernels compilation problem on non-Ampere GPUs. (#5333)
|
6 months ago |
launcher
|
3f875d9519
add device config env for the accelerator (#5396)
|
6 months ago |
linear
|
5e6c9b9311
OptimizedLinear implementation (#5355)
|
6 months ago |
model_implementations
|
d5a7c1e0b4
Capture short kernel sequences to graph (#4318)
|
10 months ago |
module_inject
|
8d98e17140
Enable mixtral 8x7b autotp (#5257)
|
6 months ago |
moe
|
08e0733e4a
Support MoE for pipeline models (#5338)
|
6 months ago |
monitor
|
604d701e35
Introduce pydantic_v1 compatibility module for pydantic>=2.0.0 support (#4407)
|
1 year ago |
nebula
|
cd4e473ee6
fix typo with deepspeed/ (#3547)
|
1 year ago |
ops
|
5e6c9b9311
OptimizedLinear implementation (#5355)
|
6 months ago |
pipe
|
b361c72761
Update DeepSpeed copyright license to Apache 2.0 (#3111)
|
1 year ago |
profiling
|
98c96e790b
Update flops profiler to handle attn and __matmul__ (#4724)
|
8 months ago |
runtime
|
c66bc4269e
set the default to use set_to_none for clearing gradients in BF16 optimizer. (#5434)
|
6 months ago |
sequence
|
2afa1c7f2f
Communication Optimization for Large-Scale Training (#4695)
|
11 months ago |
utils
|
6dcced1d5c
Cleanup required_torch_version code and references. (#5370)
|
6 months ago |
__init__.py
|
dc3554f832
Add `distributed_port` for `deepspeed.initialize` (#5260)
|
6 months ago |
accelerator
|
9548d48f48
Abstract accelerator (step 2) (#2560)
|
1 year ago |
constants.py
|
706a72562a
Allow env var for timeout (#4405)
|
1 year ago |
env_report.py
|
c08e69f212
Make op builder detection adapt to accelerator change (#5206)
|
7 months ago |
git_version_info.py
|
c08e69f212
Make op builder detection adapt to accelerator change (#5206)
|
7 months ago |
pydantic_v1.py
|
604d701e35
Introduce pydantic_v1 compatibility module for pydantic>=2.0.0 support (#4407)
|
1 year ago |