Stas Bekman c6d664849b [DeepSpeed] ZeRO Stage 3 (#10753) ...

* synced gpus

* fix

* fix

* need to use t5-small for quality tests

* notes

* complete merge

* fix a disappearing std stream problem

* start zero3 tests

* wip

* tune params

* sorting out the pre-trained model loading

* reworking generate loop wip

* wip

* style

* fix tests

* split the tests

* refactor tests

* wip

* parameterized

* fix

* workout the resume from non-ds checkpoint pass + test

* cleanup

* remove no longer needed code

* split getter/setter functions

* complete the docs

* suggestions

* gpus and their compute capabilities link

* Apply suggestions from code review

Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

* style

* remove invalid paramgd

* automatically configure zero3 params that rely on hidden size

* make _get_resized_embeddings zero3-aware

* add test exercising resize_token_embeddings()

* add docstring

Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

2021-04-08 09:53:01 -07:00

..

_static

Development on v4.6.0dev0

2021-04-06 12:53:25 -04:00

imgs

[Templates] Add template "call-for-model" markdown and "call-for-big-bird" markdown (#9921)

2021-02-05 15:47:54 +03:00

internal

Instantiate model only once in pipeline (#10888)

2021-03-29 10:39:14 -04:00

main_classes

[DeepSpeed] ZeRO Stage 3 (#10753)

2021-04-08 09:53:01 -07:00

model_doc

Typo fix of the name of BertLMHeadModel in BERT doc (#11133)

2021-04-08 08:22:58 -04:00

add_new_model.rst

Add new model docs (#9667)

2021-02-01 17:55:10 +03:00

benchmarks.rst

Make doc styler detect lists on rst (#9488)

2021-01-11 08:53:41 -05:00

bertology.rst

Fix documentation links always pointing to master. (#9217)

2021-01-05 06:18:48 -05:00

community.md

add notebook (#10995)

2021-03-31 17:00:56 +03:00

conf.py

added social thumbnail for docs (#11083)

2021-04-06 14:56:18 +02:00

contributing.md

Update installation page and add contributing to the doc (#5084)

2020-06-17 14:01:10 -04:00

converting_tensorflow_models.rst

Fix broken links in the converting tf ckpt document (#9791)

2021-01-26 03:37:57 -05:00

custom_datasets.rst

Rename NLP library to Datasets library (#10920)

2021-03-26 08:07:59 -04:00

examples.md

per_device instead of per_gpu/error thrown when argument unknown (#4618)

2020-05-27 11:36:55 -04:00

fast_tokenizers.rst

Documentation about loading a fast tokenizer within Transformers (#11029)

2021-04-05 10:51:16 -04:00

favicon.ico

Adding usage examples for common tasks (#2850)

2020-02-25 13:48:24 -05:00

glossary.rst

Adds terms to Glossary (#10443)

2021-02-28 08:27:54 -05:00

index.rst

Documentation about loading a fast tokenizer within Transformers (#11029)

2021-04-05 10:51:16 -04:00

installation.md

[doc] no more bucket

2021-04-01 14:25:47 -04:00

migration.md

Copyright (#8970)

2020-12-07 18:36:34 -05:00

model_sharing.rst

[doc] nested markup is invalid in rst (#9898)

2021-01-30 09:59:19 -05:00

model_summary.rst

ConvBERT Model (#9717)

2021-01-27 03:20:09 -05:00

multilingual.rst

Fix documentation links always pointing to master. (#9217)

2021-01-05 06:18:48 -05:00

notebooks.md

Update notebooks (#3620)

2020-04-06 14:32:39 -04:00

perplexity.rst

minor typo fix

2021-04-01 11:58:37 -06:00

philosophy.rst

Minor documentation revisions from copyediting (#9266)

2020-12-23 10:15:49 -05:00

preprocessing.rst

Minor documentation revisions from copyediting (#9266)

2020-12-23 10:15:49 -05:00

pretrained_models.rst

GPT Neo few fixes (#10968)

2021-03-30 11:15:55 -04:00

quicktour.rst

Minor documentation revisions from copyediting (#9266)

2020-12-23 10:15:49 -05:00

sagemaker.md

added new notebook and merge of trainer (#11015)

2021-04-01 23:13:47 +02:00

serialization.rst

Copyright (#8970)

2020-12-07 18:36:34 -05:00

task_summary.rst

Remove unnecessary space (#11060)

2021-04-05 09:36:20 -04:00

testing.rst

[doc] [testing] extend the pytest -k section with more examples (#10761)

2021-03-17 09:23:38 -04:00

tokenizer_summary.rst

Minor documentation revisions from copyediting (#9266)

2020-12-23 10:15:49 -05:00

training.rst

[trainer] deepspeed integration (#9211)

2021-01-12 19:05:18 -08:00