GitLab

* [feat] sync adascale from internal repo

- tbd

testing: tbd

* Update argument document of __init__

* update documentation around set_num_gradients_to_accumulate

* added checking code for proper API calling places

* rename internal APIs to make them internal

* updated changelog

* added support for add_param_group and its unit test

* added unit test for set_num_gradients_to_accumulate

* added debias_ewma unit test

* fixed test_set_num_gradients_to_accumulate (need zero_grad() call)

* added missing zero_grad() to test_lr_scheduler

* fixed test_add_param_group with respect to optim.zero_grad()

* added test_gradient_value

* added test_scale_not_equal_default for scale != world_size * grad_accum

* added test_unhook()

* removed print statements

* fixed a typo

* addressed Ben's comment

* fix typo, backend for CPU test

* Add function to add handle for sync BN
* Add test to ensure batch norm handles have been added

* catching properly a given test failing if not enough gpus

author: Joshua Meier

* file based dist init
* nicer handling of broken world sizes vs. number of available GPUs, do not break but warn out

* fix, one liner

* adjust so that frozen trunks get spread still, even if this should have little consequences

* removing dead code, hopeful unit test fix

* now with some linting..

* adding a proper unit test case

[OSS] Getting rid of the "should bucket" hash table, just use a list + non trainable params fix (#259)

* Getting rid of the "should bucket" hash table, just use a list
Properly handle all params, with or without requires_grad

* make sure that this case is unit tested

* [doc]: AdaScale example and notes

* formatted notes correctly as suggested by Benjamin

* added feature and unit test to make sure lr_scheduler works

* update the example with lr_scheduler

* fixed doc with "make html"

* addressed Mike's suggestions

* better ddp adascale tests

* make sure the single node test use the same test cases and expected gains

* added unit test that covers smoothing factor

- tested by re-introducing the bug and see the test fail as expected.

* unit test checking ddp and sharded_ddp equivalence, reproducing the issue that Sean spotted
* fixing the issue, not counting requests in flight properly
* adding a multiple optimizers case

* proper unit testing, but no other solution than disabling bucketing for now, couple of options tested do not work

* added AdaScale to README

* [adascale] added gradient accumulation

- added gradient accumulation
- tested with cifar full trainings with different value of accumulation
and verified the full accuracy is obtained
- also removed the patch optimize flag until we need it

* [adascale] adding pytest

- added basic and ddp tests and grad_accum
- closes #195

* added changelog

* added ddp grad_accum test

* moved ddp and non-ddp tests into separate files

* added checkpoint test

* more doc

* addressed Mike's comments

* fallback on internal pytorch numbering

* rewrite using autograd and Variable execution queue to make the reduce automatic
* share buckets with OSS to remove duplication
* some speed still likely on the table since the speed vs. bucketing does not match expectations, could be a follow up

add a clip gradients util, equivalent to torch's but aware of the sharded states. Add a corresponding unit test

Adds support for:
* Reused layers (e.g. for weight sharing)
* Lazily-constructed layers
* Single-process control via PipeRPCWrapper
* PipelineStyle.AsyncScheudle, which lays the foundation for asynchronous pipeline work by introducing an event loop for each rank/worker to process either activations or gradients as they arrive

Also added examples for multi-process and PipeRPCWrapper

- fixed typing
- make it run less often to reduce CI time

testing: run it in a loop make sure it is run in the right frequency.

The expert annotation is used by clip_grads and DDP.

* fixing the issue wrt Apex, validated with Latte, Classy would need another pass

* new unit test to catch rank issues in OSS

Currently only implemented for a single process and expert.

* Add unittest for checkpoint & DDP

- this change adds test cases to reproduce the error with checkpoint & DDP
- mandeep mentioned that there is also deadlock in this case, but this
  change doesn't cover that.
- we cover cases where weight sharing is OK
- however, same module multiple checkpoint or find_unused_parameters are
  both not OK

* added norm checks

Same bucketing strategy for OSS and SDP:
sort everything ahead of time, per rank and per size, smaller tensors first. Bucket the smallest elements in a fixed buffer, send async, then send all the others async, and get back to the bucket. Once done then scatter the contents if needed