.. |
accelerator
|
b20c46745b
add missing methods to MPS_Accelerator (#5134)
|
8 months ago |
autotuning
|
b361c72761
Update DeepSpeed copyright license to Apache 2.0 (#3111)
|
1 year ago |
checkpoint
|
774b897736
fix the missing argument in test and typo (#5730)
|
3 months ago |
comm
|
19da95f783
[CPU] add fp16 support to shm inference_all_reduce (#5669)
|
3 months ago |
compression
|
6dcced1d5c
Cleanup required_torch_version code and references. (#5370)
|
6 months ago |
elasticity
|
8e4f6e48db
Skip the UT cases that use unimplemented op builders. (#5372)
|
5 months ago |
hybrid_engine
|
f69f8840fc
Removal of cuda hardcoded string with get_device function (#5351)
|
6 months ago |
inference
|
1a45bd8e8c
Lock cache file of HF model list (#6628)
|
6 days ago |
launcher
|
13c16c9562
Accept btl_tcp_if_include option through launcher_args (#6613)
|
1 week ago |
linear
|
6e5d58d248
OptimizedLinear updates (#5791)
|
2 months ago |
model_parallelism
|
6dcced1d5c
Cleanup required_torch_version code and references. (#5370)
|
6 months ago |
moe
|
9a3ede7079
add moe topk(k>2) gate support (#5881)
|
2 months ago |
monitor
|
488a823f64
New integration - CometMonitor (#5466)
|
5 months ago |
ops
|
a1f98bdc70
AIO CPU Locked Tensor (#6592)
|
1 week ago |
pipe
|
7ddc3b01dd
Fix pipeline module evaluation when contiguous activation checkpointing is enabled (#3005)
|
1 year ago |
profiling
|
6dcced1d5c
Cleanup required_torch_version code and references. (#5370)
|
6 months ago |
runtime
|
85b7469ea0
Add first Step in LR Schedulers (#6597)
|
1 week ago |
sequence_parallelism
|
8b191d7ccf
Long sequence parallelism (Ulysses) integration with HuggingFace (#5774)
|
2 months ago |
utils
|
08e0733e4a
Support MoE for pipeline models (#5338)
|
6 months ago |
__init__.py
|
b361c72761
Update DeepSpeed copyright license to Apache 2.0 (#3111)
|
1 year ago |
alexnet_model.py
|
9bc4cd01b7
Store/Load CIFAR from local/offline (#6390)
|
1 month ago |
common.py
|
c9fc34a4be
Use file store for tests (#6632)
|
4 days ago |
ds_batch_config.json
|
ff42743865
Refactor remaining distributed tests (#2216)
|
2 years ago |
gpt2-merges.txt
|
ff42743865
Refactor remaining distributed tests (#2216)
|
2 years ago |
gpt2-vocab.json
|
ff42743865
Refactor remaining distributed tests (#2216)
|
2 years ago |
megatron_model.py
|
4b35833379
Revert "Update megatron GPT2Model"
|
1 year ago |
modeling.py
|
180dd39714
Clean up modeling code (#4320)
|
1 year ago |
modelingpreln.py
|
180dd39714
Clean up modeling code (#4320)
|
1 year ago |
multi_output_model.py
|
c08e69f212
Make op builder detection adapt to accelerator change (#5206)
|
7 months ago |
simple_model.py
|
c08e69f212
Make op builder detection adapt to accelerator change (#5206)
|
7 months ago |
util.py
|
1ab1928d79
Enable dynamic shapes for pipeline parallel engine inputs (#5481)
|
2 months ago |