Skip to content

Conversation

@jeffra
Copy link
Collaborator

@jeffra jeffra commented May 19, 2020

No description provided.

jeffra and others added 30 commits May 15, 2020 18:01
Updates for ZeRO stage 2 + ZeRO stage 1 w. RS

Co-authored-by: Tunji Ruwase <olruwase@microsoft.com>
Co-authored-by: Samyam Rajbhandari <samyamr@microsoft.com>
* Fixed
    - tests/unit/test_fp16.py::test_zero_allow_untested_optimizer
    - tests/unit/test_fp16.py::test_adam_fp16_onecycle_compatibility
    - tests/unit/test_fp16.py::test_adam_fp16_zero_stage_1_onecycle_compatibility
    - tests/unit/test_fp16.py::test_adam_fp16_zero_stage_2_onecycle_compatibility
    - tests/unit/test_checkpointing.py::test_checkpoint_zero_optimizer[2]

* Still broken
    - tests/unit/test_fp16.py::test_zero_empty_partition[1,2]
    - tests/unit/test_fp16.py::test_zero_static_scale[1,2]
    - tests/unit/test_checkpointing.py::test_checkpoint_zero_optimizer[1]
…nting. Toggle different optimizations in deepspeed.checkpointing via deepspeed_config json
…d json config documentation for ZeRO-2 \ Added API webpage for DeepSpeed Activation Checkpointing \ Updated Megatron Tutorial \ Added json config documnetation for activation checkpointing
ShadenSmith and others added 27 commits May 15, 2020 18:11
Co-authored-by: yuxionghe <yuxhe@microsoft.com>
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>
* finalizing test for base deepspeed and deepspeed with ZeRO

* applying the comment (thanks Jeff); fixed formatting

* adding back old version of zero optimization config setup

* adding unit tests and applying comments + formatting
@jeffra jeffra merged commit f2ac7ea into master May 19, 2020
@jeffra jeffra deleted the zero2-staging branch September 15, 2020 20:29
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

7 participants