GitLab

Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

* Fix mis-aligned-grad

When a parameter is not divisible by world size, the partitioned gradients are mis-aligned due to incorrect padding handling. This PR should fix for that.

* Formatting fix

* Adding static_scale test back for Z3, and also changing hidden size to be not divisile by world_size

* also removing alignment from flat fp16 buffers

* Testing for hidden dim alignment

* inference hook fix

* Update stage3.py

* formatting

* [bug-fix] move params to gpu if offload params is turned off
Co-authored-by: Samyam Rajbhandari <samyamr@microsoft.com>
Co-authored-by: Shaden Smith <Shaden.Smith@microsoft.com>
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

* Squash stage3 v1 (#146)
Co-authored-by: Samyam <samyamr@microsoft.com>
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>
Co-authored-by: Samyam Rajbhandari <samyamr@microsoft.com>
Co-authored-by: Olatunji Ruwase <olruwase@microsoft.com>
Co-authored-by: Shaden Smith <Shaden.Smith@microsoft.com>
Co-authored-by: Shaden Smith <ShadenTSmith@gmail.com>
Co-authored-by: eltonzheng <eltonz@microsoft.com>

* Fix correctness bug (#147)

* formatting fix (#150)

* stage3 bugfix (API) update and simplified FP16 Z3 tests (#151)

* fp16 Z3 API update and bugfix

* revert debug change

* ZeRO-3 detach and race condition bugfixes (#149)

* trying out ZeRO-3 race condition fix

* CUDA sync instead of stream

* reduction stream sync

* remove commented code

* Fix optimizer state_dict KeyError (#148)
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

* fix for smaller SGS sizes, ensures each grad is backed by unique tensors (#152)

* Simplifying the logic for getting averaged gradients (#153)

* skip for now

* Z3 Docs redux (#154)

* removing some TODOs and commented code (#155)

* New Z3 defaults (#156)
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

* formatting

* megatron external params
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>
Co-authored-by: Olatunji Ruwase <olruwase@microsoft.com>
Co-authored-by: Shaden Smith <Shaden.Smith@microsoft.com>
Co-authored-by: Shaden Smith <ShadenTSmith@gmail.com>
Co-authored-by: eltonzheng <eltonz@microsoft.com>

* Activation checkpoint support for non tensor input/output

* Format fixes

* Address PR comments; Add ordering edge case tests

* work on flops profiler tutorial

* update flops profiler tutorial

* add flops profiler tutorial and fix names

* work on flops profiler tutorial

* update flops profiler tutorial

* add flops profiler tutorial and fix names

* fix tailing ws

* fix names

* remove multistep profiling and update docs

* fix cases where functionals and submodules coexist in a parent module, update readme

* fix typo

* always invoke post hook function

* fix module flops sum and update tests

* update tutorial

Co-authored-by: Cheng Li <pistasable@gmail.com>
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

Special thanks to @g-karthik for tracking this issue down.

* Add Linear warmup+decay lr schedule
Update lr schedule unit tests

* LR scheduler unit tests for LR Range Test and 1Cycle

* Disable yapf to preserve parameterizaton

* Disable test_pipe.py for CI debugging

* Disable test_lr_scheduler for CI debugging

* Disable test_lr_scheduler for CI debugging

* Enable all unit tests for CI debugging
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

Co-authored-by: Reza Yazdani <reyazda@microsoft.com>
Co-authored-by: Olatunji Ruwase <olruwase@microsoft.com>

Allow DeepSpeed models to be initialized with optimizer=None
Co-authored-by: Shaden Smith <Shaden.Smith@microsoft.com>

Co-authored-by: Samyam Rajbhandari <samyamr@microsoft.com>

Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

* supporting different hidden dimensions

* add support for larger hidden dimensions (greater than 8K)

* remove empty line

* add loop unrolling factor for dropout kernels

* update different kernels based on the reviews
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

* Use zero-tensors for missing gradients to avoid size mismatch

* Unit test for unbalanced gradients in ZeRO

* Formatting fixes

* zero-1 memory fix

* auto-tune max elems per comm to reduce padding/comm intervals

* clean-up and added previously missing reduction options

* fix testing backing to work with torch1.7

* Fix layout bug in ZeRO Stage 1 checkpoint logic
Add elastic checkpoint option for ZeRO stage 1, default to True

* Format fixes

Co-authored-by: Shaden Smith <Shaden.Smith@microsoft.com>
Co-authored-by: Reza Yazdani <reyazda@microsoft.com>

* Progressive layer dropping docs (#499)

* test

* Adding tutorial and news page for pld

* updating the tutorial and posts of PLD

* update the finetune tutorial

* Update PLD tutorial (#512)

* Update installation instructions

* Format fix

* ZeRO tutorial

* Format fixes

* ZeRO-Offload

* ZeRO and ZeRO-Offload tutorials

* Update navigation page

* Format fixes

* Add yuxhe feedback

* Fix blog post link

* Fix OneBit-Adam link
Tweak scheduler example

* Fix date link

* Add DeepSpeed_Adam

* Add PLD tutorial to navigation
Co-authored-by: Shaden Smith <Shaden.Smith@microsoft.com>
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

* updating the pld docs

* DeepSpeed implementation of PLD (#508)

* DeepSpeed implementation of PLD

* Format fixes

* Formatting fixes

* Fix broken url

* Address PR feedback

* Bump DSE
Co-authored-by: Minjia Zhang <33713995+minjiaz@users.noreply.github.com>
Co-authored-by: Shaden Smith <Shaden.Smith@microsoft.com>
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>
Co-authored-by: Minjia Zhang <minjiaz@microsoft.com>

* add adamW to CPU-ADAM implementation

* supporting cpu-adam optimizer for zero-offload on deepspeed side

* bump DSE to match cpu-adam updates
Co-authored-by: Jeff Rasley <jerasley@microsoft.com>

* Disable default installation of CPU Adam

* Handle cpufeature import/use errors separately

Co-authored-by: Conglong Li <conglong.li@gmail.com>

This reverts commit 01b6e27e

.
Co-authored-by: Shaden Smith <ShadenTSmith@gmail.com>