Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[WIP] [doc] performance/scalability revamp #15723

Merged
merged 26 commits into from
May 16, 2022
Merged

[WIP] [doc] performance/scalability revamp #15723

merged 26 commits into from
May 16, 2022

Conversation

stas00
Copy link
Contributor

@stas00 stas00 commented Feb 18, 2022

moved from #15213 so that we get the doc generator working

XXX: The previous PR has comments/suggestions that need to be integrated here


@lvwerra and I are working on a massive performance/scalability docs revamp:

So the rough plan is to make custom plans for each of the combinations [inference|training] * [1 gpu|many gpus|cpu] so that it's very easy for the user to follow the instructions that are specific to their needs.

So the proposed doc layout is:

  • performance.mdx (main entry point)
  • perf_infer.mdx
    • perf_infer_cpu.mdx
    • perf_infer_gpu_many.mdx
    • perf_infer_gpu_one.mdx
  • perf_train.mdx
    • perf_train_gpu_many.mdx
    • perf_train_gpu_one.mdx
  • scalability.mdx (rename from parallelism.mdx) (XXX: to do)

See the PR's changes for a rough layout of the content.

One big question is this: At the moment everything is pytorch-centric, as we don't have any info on tf/flax. Down the road we will either inject tf/flax-specific instructions into the current docs, or perhaps it'd be better to have dedicated docs for pt/tf/flax. It'd help a lot to decide ahead of time to avoid document renaming and potentially breaking links. If we plan to have these PT-specific perhaps let's embed _pt in the filenames?

@lvwerra

@HuggingFaceDocBuilder
Copy link

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint.

@lvwerra
Copy link
Member

lvwerra commented Mar 15, 2022

Hi @stas00, first stab at the single GPU section. The text is still WIP but you could have look at the sections and rough content to see if you agree.

Looking at the docs as a whole again here a few thoughts:

  • Instead of dividing the whole section into speed and memory related techniques I thought an alternative could be to have a summary table at the beginning or end where the memory and speed impact of all methods is described. Could also help a user navigate the doc better.
  • In the previous PR we discussed where to discuss common concepts (first time they occurred vs. in the performance docs). I am currently leaning towards the former. The main thing that's currently in the performance.mdx and I am not sure, yet, where I would put it is the hardware discussion which is quite different to everything else we are doing which makes me think that maybe we can create a dedicated document hardware.mdx where we could also expand a bit to other accelerators (TPU/IPUs for example).
  • If we take that road the performance.mdx doc is again quite empty and I think we could move the content from perf_infer.mdx and perf_train.mdx to this document and then have a shared entry point and remove perf_infer.mdx and perf_train.mdx.

If you are happy with the structure I could start outlining the performance.mdx and polishing the single-GPU section while you could for example work on the multi-GPU training content.

What do you think?

@stas00
Copy link
Contributor Author

stas00 commented Mar 17, 2022

Thank you for taking a stab at redesign, @lvwerra

It's very clear that we have fundamentally conflicting visions of how the performance documentation should be laid out. So it'd be wasteful to spend energy to try to come up with something that resonates with both of us.

I don't think this project can have 2 masterminds, and since I'm currently busy with BigScience training and we don't want this to fall between the cracks, I propose that if this works for you, please take over and proceed in an unimpeded way that sounds good to you, @lvwerra.

Once you have completed the restructuring and added what you know and you want me to fill in some gaps that I hopefully know to fill in please let me know and I will do so. or alternatively do tell me what sections are missing and I will work on those in free format and you can then adapt them to the structure you end up choosing.

@lvwerra lvwerra requested review from sgugger and LysandreJik April 13, 2022 09:59
@lvwerra
Copy link
Member

lvwerra commented Apr 13, 2022

Hi @LysandreJik and @sgugger,

I revamped the documentation a bit and reduced the scope from all the documents @stas00 lined out to just the performance.mdx and the training on single/multi-GPU so we can converge on those before adding the others, which also prevents a monster PR. For some reason the preview of the docs still does not work - if you have any idea, let me know.

To guide you a bit in the review:

  • performance.mdx should be the entry point and give an overview
  • perf_train_gpu_single.mdx shows all the tricks to efficiently train large models on a single GPU (gradient accumulation/checkpointing, optimizers, DeepSpeed Zero, data loaders etc.)
  • perf_train_gpu_single.mdx shows all the tricks to efficiently train large models on a single GPU (mostly parallelism strategies: data, tensor, pipeline parallelism)
  • perf_hardware.mdx is where I put some tips and tricks about custom hardware setups

In subsequent PRs we can:

  • add more TensorFlow examples
  • add TPU training guide
  • add inference side
  • add guides for train/inference on specialized hardware (Optimum)

This is the plan for the full documentation:

perf_overview

Let me know what you think about style and content of the documentation. There are still a few rough edges but I think it is good for a first review.

@sgugger
Copy link
Collaborator

sgugger commented Apr 13, 2022

The preview can't work if you don't rebase on master to have the new doc structure (everything nested in an "en" subfolder)

@lvwerra lvwerra force-pushed the doc-perf-revamp-2 branch from 5c72639 to 5fb8e15 Compare April 13, 2022 12:06
@lvwerra lvwerra force-pushed the doc-perf-revamp-2 branch from 5fb8e15 to cd93337 Compare April 13, 2022 12:30
@HuggingFaceDocBuilderDev
Copy link

HuggingFaceDocBuilderDev commented Apr 13, 2022

The documentation is not available anymore as the PR was closed or merged.

@lvwerra
Copy link
Member

lvwerra commented Apr 13, 2022

Thank you @sgugger it worked!

Comment on lines 80 to 81
- local: fast_tokenizers
title: "Using tokenizers from 🤗 Tokenizers"
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This change does not seem to be linked to this PR.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good catch!must have happened during rebase - I'll remove it.

docs/source/en/perf_hardware.mdx Outdated Show resolved Hide resolved
docs/source/en/perf_train_gpu_many.mdx Outdated Show resolved Hide resolved
docs/source/en/perf_train_gpu_many.mdx Show resolved Hide resolved
docs/source/en/perf_train_gpu_one.mdx Outdated Show resolved Hide resolved
docs/source/en/perf_train_gpu_one.mdx Outdated Show resolved Hide resolved
docs/source/en/perf_train_gpu_one.mdx Outdated Show resolved Hide resolved
docs/source/en/perf_train_gpu_one.mdx Outdated Show resolved Hide resolved
docs/source/en/perf_train_gpu_one.mdx Outdated Show resolved Hide resolved
Comment on lines +569 to +570
### `_multi_tensor`
pytorch-nightly introduced `torch.optim._multi_tensor` which should significantly speed up the optimizers for situations with lots of small feature tensors. It should eventually become the default, but if you want to experiment with it sooner and don't mind using the bleed-edge, see: https://github.com/huggingface/transformers/issues/9965
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is one year old, is it really the bleeding edge?

Copy link
Contributor Author

@stas00 stas00 Apr 13, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I did spend time investigating this a few months ago, you can see the details here: pytorch/pytorch#71274

the pytorch devs swear it's supposed to be faster, but I wasn't able to see any speed improvement. tldr: pytorch/pytorch#71274 (comment)

I have a separate PR that expands the optimizers section: #14708 and I was thinking to add it after the dust settles, but I suppose it might be better to put it into the "fodder" now?

Copy link
Collaborator

@sgugger sgugger Apr 13, 2022

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

As long as it's not forgotten here I'm fine with both solutions.

lvwerra and others added 2 commits April 14, 2022 10:23
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
leandro and others added 4 commits April 14, 2022 10:33
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
@lvwerra lvwerra marked this pull request as ready for review April 19, 2022 11:59
@stas00
Copy link
Contributor Author

stas00 commented Apr 20, 2022

@lvwerra, I rebased and solved the conflict introduced by #16860 - fixing it at its new file the content was moved to.

@lvwerra
Copy link
Member

lvwerra commented Apr 22, 2022

@stas00 thank you!

@sgugger @LysandreJik would you like to take a look at the PR? I integrated the previous suggestion.

Copy link
Member

@LysandreJik LysandreJik left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Impressive refactor!

@lvwerra
Copy link
Member

lvwerra commented May 6, 2022

@sgugger are you happy to merge this?

docs/source/en/_toctree.yml Outdated Show resolved Hide resolved
@sgugger
Copy link
Collaborator

sgugger commented May 9, 2022

Yes, LGTM :-)

@lvwerra lvwerra merged commit 71abd3a into main May 16, 2022
@lvwerra lvwerra deleted the doc-perf-revamp-2 branch May 16, 2022 11:36
ArthurZucker pushed a commit to ArthurZucker/transformers that referenced this pull request May 17, 2022
* [doc] performance/scalability revamp

* link the new docs

* no :

* mixed precision

* work on the first doc

* expand the main doc

* Trigger CI

* style

* revamp single GPU training section

* work on training performance

* remove files not used anymore or will be added later

* final touches

* fix rebase

* Add hardware section to toctree

* fix toctree again

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* remove `fast_tokenizers` entry that was copied in rebase

* add warning about DP vs DDP

* remove todo

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* fix missing closure of codeblock

* Update docs/source/en/perf_train_gpu_many.mdx

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* sync with huggingface#16860

* update toc

Co-authored-by: leandro <leandro.vonwerra@spoud.io>
Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Narsil pushed a commit to Narsil/transformers that referenced this pull request May 30, 2022
fix tokenizer autodoc

fix minor CI issues

fix minor CI issues

fix minor CI issues

fix style issue

fix minor import issues

fix few issues

remove def main on the test

add require torch

replace decorator with 'with'

fix style

change to bloom

add quick fix tokenizer

fix tokenizer file

fix tokenizer

- merge tests
- small fixes

fix import issue

add bloom to readme

fix consistency

Update docs/source/en/model_doc/bloom.mdx

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

Apply suggestions from code review

fix comment issues on file headers

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

fix doc issue

small fix - modeling test

some changes

- refactor some code
- taking into account reviews
- more tests should pass
- removed pruning tests

remove useless division

more tests should pass

more tests should pass

more tests should pass

let's try this one

-add alibi offset
- remove all permutes to make the grad operations work
- finger crossed

Update data2vec.mdx to include a Colab Notebook link (that shows fine-tuning) (huggingface#17194)

* Update data2vec.mdx

* Update data2vec.mdx

* Update docs/source/en/model_doc/data2vec.mdx

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

Dev version

Add test to ensure models can take int64 inputs (huggingface#17210)

* Add test to ensure models can take int64 inputs

* is_integer is an attribute, not a method

* Fix test when some inputs aren't tensors

* Add casts to blenderbot and blenderbot-small

* Add casts to the other failing models

Fix dependency table

update BART docs (huggingface#17212)

Black preview (huggingface#17217)

* Black preview

* Fixup too!

* Fix check copies

* Use the same version as the CI

* Bump black

Fix typo in bug report template (huggingface#17178)

* Fix typo

* Force rerun workflows

Co-authored-by: Felix Marty <felix@huggingface.co>

Added translation of installation.mdx to Portuguese Issue huggingface#16824 (huggingface#16979)

* Added translation of installation.mdx to Portuguese, as well
as default templates of _toctree.yml and _config.py

* [ build_documentation.yml ] - Updated doc_builder to build
documentation in Portuguese.
[ pipeline_tutorial.mdx ] - Created translation for the pipeline_tutorial.mdx.

* [ build_pr_documentation.yml ] - Added pt language to pr_documentation builder.

[ pipeline_tutorial.mdx ] - Grammar changes.

* [ accelerate.mdx ] - Translated to Portuguese the acceleration tutorial.

* [ multilingual.mdx ] - Added portuguese translation for multilingual tutorial.

[ training.mdx ] - Added portuguese translation for training tutorial.

* [ preprocessing.mdx ] - WIP

* Update _toctree.yml

* Adding Pré-processamento to _toctree.yml

* Update accelerate.mdx

* Nits and eliminate preprocessing file while it is ready

Co-authored-by: Omar U. Espejel <espejelomar@gmail.com>

OPT-fix (huggingface#17229)

* try fixes

* Revert "try fixes"

This reverts commit a8ad75e.

* add correct shape

* add correct path

OPT - fix docstring and improve tests slighly (huggingface#17228)

* correct some stuff

* fix doc tests

* make style

Update self-push workflow (huggingface#17177)

* update push ci

* install git-python

* update comment

* update deepspeed jobs

* fix report

* skip 2 more tests that require fairscale

* Fix changes in test_fetcher.py (to deal with `setup.py` is changed)

* set RUN_PT_TF_CROSS_TESTS=1 and final clean-up

* remove SIGOPT_API_TOKEN

* remove echo "$matrix_folders"

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

fix --gpus option for docker (huggingface#17235)

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

Handle copyright in add-new-model-like (huggingface#17218)

Fix Trainer for Datasets that don't have dict items (huggingface#17239)

install dev. version of accelerate (huggingface#17243)

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

Fix push CI channel (huggingface#17242)

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

Add PR title to push CI report (huggingface#17246)

* add PR title to push CI report

* add link

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

[ fast_tokenizers.mdx ] - Added translation to portuguese to tutorial (huggingface#17076)

* [ fast_tokenizers.mdx ] - Added translation to portuguese to tutorial

* Delete docs/source/pt-br directory

* [ fast_tokenizers.mdx ] - Continuing work on file

* [ fast_tokenizers.mdx ] - Continuing work on file

* Add fast tokenizers to _toctree.yml

* Eliminated config and toctree.yml

* Nits in fast_tokenizers.mdx

Co-authored-by: Omar U. Espejel <espejelomar@gmail.com>

Translated version of model_sharing.mdx doc to spanish (huggingface#16184)

* Translated version of model_sharing to spanish

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Update docs/source_es/model_sharing.mdx

* Addind model sharing to _toctree.yml

Co-authored-by: Omar U. Espejel <espejelomar@gmail.com>

Guide to create custom models in Spanish (huggingface#17158)

* file copied and toctree updated

* Intro and configuration translated

* model section translated

* enter hotfix

* Translation over, correction pending

* Typos and corrections

* Update docs/source/es/create_a_model.mdx

Co-authored-by: Omar U. Espejel <espejelomar@gmail.com>

* Update docs/source/es/create_a_model.mdx

Co-authored-by: Omar U. Espejel <espejelomar@gmail.com>

* Update docs/source/es/create_a_model.mdx

Co-authored-by: Omar U. Espejel <espejelomar@gmail.com>

* Update docs/source/es/create_a_model.mdx

Co-authored-by: Omar U. Espejel <espejelomar@gmail.com>

Co-authored-by: Omar U. Espejel <espejelomar@gmail.com>

Fix obvious typos in flax decoder impl (huggingface#17279)

Change config.encoder_ffn_dim -> config.decoder_ffn_dim for decoder.

TF - Fix convnext classification example (huggingface#17261)

[WIP] [doc] performance/scalability revamp (huggingface#15723)

* [doc] performance/scalability revamp

* link the new docs

* no :

* mixed precision

* work on the first doc

* expand the main doc

* Trigger CI

* style

* revamp single GPU training section

* work on training performance

* remove files not used anymore or will be added later

* final touches

* fix rebase

* Add hardware section to toctree

* fix toctree again

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* remove `fast_tokenizers` entry that was copied in rebase

* add warning about DP vs DDP

* remove todo

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* fix missing closure of codeblock

* Update docs/source/en/perf_train_gpu_many.mdx

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* sync with huggingface#16860

* update toc

Co-authored-by: leandro <leandro.vonwerra@spoud.io>
Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

fixed bug in run_mlm_flax_stream.py (huggingface#17203)

* fixed bug run_mlm_flax_stream.py

Fixed bug caused by an update to tokenizer keys introduced in recent transformers versions (between `4.6.2` and `4.18.0`) where additional keys were introduced to the tokenizer output.

* Update run_mlm_flax_stream.py

* adding missing paranthesis

* formatted to black

* remove cols from dataset instead

* reformat to black

* moved rem. columns to map

* formatted to black

Co-authored-by: KennethEnevoldsen <kennethcenevolsen@gmail.com>

 Updated checkpoint support for Sagemaker Model Parallel (huggingface#17219)

* adding partial checkpoint support for optimizer state

* formatted trainer.py

* Refactoring based on comments

* reformatting

* Update src/transformers/trainer.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/trainer.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/trainer.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

Co-authored-by: Cavdar <dcavdar@a07817b12d7e.ant.amazon.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

Update codeparrot data preprocessing (huggingface#16944)

* add new preprocessing arguments

* add new filters

* add new filters to readme

* fix config and test count, update function names and docstrings

* reformat code

* update readme

* Update readme

* rename config_test filter

Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* rename few_assignments filter

Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* rename tokenizer in arguments

Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* rename functions and add limit_line argument for config_test filter

* update threshold for config_test filter

Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>
Co-authored-by: Loubna ben allal <loubnabenallal@gmail.com>

CodeParrot data pretokenization (huggingface#16932)

* add pretokenization arguments

* add pretokenization script

* add support for pretokenized data

* reformat code

* fix run command for training

* fix model call from config

* remove a package

* add comments on pretokenization in the readme

* remove explicit parallelization

Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* update readme

Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* update readme -remove username

Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* update readme -remove username

Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

* keep data parallelization

* reformat code

* reformat code

* update readme

* reformat code

* Update examples/research_projects/codeparrot/README.md

Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>

Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>
Co-authored-by: Loubna ben allal <loubnabenallal@gmail.com>

Remove next sentence prediction from supported ONNX tasks (huggingface#17276)

Align logits and labels in OPT (huggingface#17237)

Mlflowcallback fix nonetype error (huggingface#17171)

* Fix edge cases TypeError: 'NoneType' object is not callable

* fix style

Automatically sort auto mappings (huggingface#17250)

* Automatically sort auto mappings

* Better class extraction

* Some auto class magic

* Adapt test and underlying behavior

* Remove re-used config

* Quality

Make TrainerHyperParameterSigOptIntegrationTest slow test (huggingface#17288)

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

Better error in the Auto API when a dep is missing (huggingface#17289)

Fix FlavaForPreTrainingIntegrationTest CI test (huggingface#17232)

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

Use the PR URL in CI report (huggingface#17269)

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

logging documentation update (huggingface#17174)

* logging documentation

* style

Co-authored-by: Sander Land <sander@chatdesk.com>

docs(transformers): fix typo (huggingface#17263)

Add Tensorflow Swin model (huggingface#16988)

Co-authored-by: Matt <Rocketknight1@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

[Tests] Fix slow opt tests (huggingface#17282)

* fix opt tests

* remove unused tok

* make style

* make flake8 happy

* Update tests/models/opt/test_modeling_opt.py

Fix test_model_parallelization (huggingface#17249)

* Fix test_model_parallelization

* Modify

Add Wav2Vec2Conformer (huggingface#16812)

* save intermediate

* add wav2vec2 conformer

* add more code

* more

* first test passes

* make all checkpoints work

* update

* up

* more clean ups

* save clean-up

* save clean-up

* save more

* remove bogus

* finalize design conformer

* remove vision

* finish all tests

* more changes

* finish code

* add doc tests

* add slow tests

* fix autoconfig test

* up

* correct docstring

* up

* update

* fix

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: Anton Lozhkov <aglozhkov@gmail.com>

* Update docs/source/en/model_doc/wav2vec2-conformer.mdx

* upload

* save copied from

* correct configs

* fix model outputs

* add to docs

* fix imports

* finish

* finish code

* correct copied from

* correct again

* correct make fix

* improve make fix copies

* save

* correct fix copy from

* correct init structure

* correct

* fix import

* apply suggestions

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: Anton Lozhkov <aglozhkov@gmail.com>

Fix missing job action button in CI report  (huggingface#17270)

* use matrix.machine_type

* fix job names used in job_link

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

Fix wrong PT/TF categories in CI report (huggingface#17272)

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

[ConvNeXT] Fix drop_path_rate (huggingface#17280)

* Fix drop_path_rate

* Fix TF's drop path rate

fix retribert's `test_torch_encode_plus_sent_to_model` (huggingface#17231)

Fix tests of mixed precision now that experimental is deprecated (huggingface#17300)

* Fix tests of mixed precision now that experimental is deprecated

* Fix mixed precision in training_args_tf.py too

Rewrite TensorFlow train_step and test_step (huggingface#17057)

* Initial commit

* Better label renaming

* Remove breakpoint before pushing (this is your job)

* Test a lot more in the Keras fit() test

* make fixup

* Clarify the case where we flatten y dicts into tensors

* Clarify the case where we flatten y dicts into tensors

* Extract label name remapping to a method

correct opt (huggingface#17301)

refactor

- refactor code
- style changes
- add new threshold for test

major changes

- change BLOOM to Bloom
- add quick doc on bloom.mdx
- move embeddings test on modeling test

modify readme

small fixes

small fix

- better threshold for a test

remove old test file from fetcher

fix small typo

major change

- change BloomLMHead to BloomForCausalLM

remove onnx config

major changes

- refactor the code
- remove asserts
- change tol for test

make style

small change

adding a slow test + commenting old ones for now

make style

Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

make style

fix duplicates

cleaning comments on config

clean a bit conversion file

refacor a bit modeling file

refactor tokenizer file

fix tokenization test issue

fix tokenization issue second try

fix tokenization issue #2

fix test issue

make style + add suggestions

change test fetcher

try this one

- slow tests should pass
- finger crossed

possible final changes

make style

try fix padding side issue

fix side

fix padding issue

fix ko-readme

fix config auto

cleaning modeling file

keep bloom in caps in ko

update config docs

remove pretraining_pp

remove model parallel

update config

- add correct config files

fix duplicates

fix fetcher

fix refactor issue

- remove divide function

try to remove alibi

small fixes

- fix alibi
- remove seq length
- refactor a bit the code

put correct values

- fix bos and eos token ids

fix attention mask loop

Co-authored-by: thomasw21 <24695242+thomasw21@users.noreply.github.com>

small fixes:

- remove skip bias add

small fixes

- fix typo in readme
- fix typos in config

small changes

- remove a test
- add reconstruction test
- change config

small changes

- change Scaled Softmax to BloomScaledSoftmax

small fixes

- fix alibi dtype

major changes

- removing explicit dtype when loading modules
- fixing test args (torch_dtype=auto)
- add dosctring

fix readmes

major changes

- now bloom supports alibi shifting
- refactor a bit the code
- better test tolerance now

refactor a bit

refactor a bit

put correct name on test

change docstring

small changes

- fix docstring modeling
- fix test tolerance

fix small nit

- take dtype from tensors in the conversion script

minor fix

- fix mdx issue

minor fix

- change config docstring

forward contrib credits from PR14084

Apply suggestions from code review

Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>

apply modifications

Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>

resolve softmax upcast

Apply suggestions from code review

Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>

Update src/transformers/models/bloom/modeling_bloom.py

Co-authored-by: Niklas Muennighoff <n.muennighoff@gmail.com>

final changes modeling

Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>

Merge commit 'd156898f3b9b2c990e5963f5030a7143d57921a2'

merge commit

Apply suggestions from code review

Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>

apply suggestions

Apply suggestions from Stas comments
Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>
elusenji pushed a commit to elusenji/transformers that referenced this pull request Jun 12, 2022
* [doc] performance/scalability revamp

* link the new docs

* no :

* mixed precision

* work on the first doc

* expand the main doc

* Trigger CI

* style

* revamp single GPU training section

* work on training performance

* remove files not used anymore or will be added later

* final touches

* fix rebase

* Add hardware section to toctree

* fix toctree again

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* remove `fast_tokenizers` entry that was copied in rebase

* add warning about DP vs DDP

* remove todo

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* fix missing closure of codeblock

* Update docs/source/en/perf_train_gpu_many.mdx

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* sync with huggingface#16860

* update toc

Co-authored-by: leandro <leandro.vonwerra@spoud.io>
Co-authored-by: Leandro von Werra <lvwerra@users.noreply.github.com>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
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.

6 participants