NielsRogge/transformers

Pull Requests Commits

julien-c committed 5 years ago

3e0f0621

Fix: one more try

julien-c committed 5 years ago

fc2a4c88

Same fix for `addcmul_`

julien-c committed 5 years ago

55bda525

Fix UserWarning: This overload of add_ is deprecated in pytorch==1.5.0

julien-c committed 5 years ago

ad02c961

[skip ci] remove local rank

julien-c committed 5 years ago

15550ce0

rerun notebook 02-transformers (#4341)

nikitajz committed 5 years ago

Verified 62427d08

Allow for None gradients in GradientAccumulator. (#4372)

jarednielsen committed 5 years ago

Verified 34706ba0

Should return overflowing information for the log (#4385)

LysandreJik committed 5 years ago

Verified edf9ac11

Attempt to unpin torch version for Github Action. (#4384)

mfuntowicz committed 5 years ago

Verified b908f2e9

[examples] Streamline doc

julien-c committed 5 years ago

af2e6bf8

p_mask in SQuAD pre-processing (#4049)

LysandreJik committed 5 years ago

Verified 7defc667

Updated ONNX notebook link in README.

mfuntowicz committed 5 years ago

84894974

Conversion script to export transformers models to ONNX IR. (#4253)

mfuntowicz committed 5 years ago

Verified db0076a9

Fix trainer evaluation (#4363)

patil-suraj committed 5 years ago

Verified 2d054801

Create README.md (#4359)

savasy committed 5 years ago

Verified 035678ef

Create README.md (#4357)

sy-wada committed 5 years ago

Verified b9c9e053

Tokenizer.batch_decode convenience method (#4159)

sshleifer committed 5 years ago

Verified 9535bf19

[tests] make pipelines tests faster with smaller models (#4238)

sshleifer committed 5 years ago

Verified 7822cd38

Fix: unpin flake8 and fix cs errors (#4367)

julien-c committed 5 years ago

Verified 448c4672

Use Filelock to ensure distributed barriers

julien-c committed 5 years ago

c547f15a

[ci skip] Pin isort

julien-c committed 5 years ago

015f7812

TPU needs a rendezvous (#4339)

LysandreJik committed 5 years ago

Verified ef46ccb0

Add image and metadata (#4345)

ViktorAlm committed 5 years ago

Verified 94cb73c2

Add link to W&B to see whole training logs (#4348)

mrm8488 committed 5 years ago

Verified a0eebdc4

Release: v2.9.1

LysandreJik committed 5 years ago

7cb203fa

[Marian Fixes] prevent predicting pad_token_id before softmax, support language codes, name multilingual models (#4290)

sshleifer committed 5 years ago

Verified 9a687ebb

[Docs, Notebook] Include generation pipeline (#4295)

patrickvonplaten committed 5 years ago

Verified 839bfaed

wrong variable name used (#4328)

elyesmanai committed 5 years ago

Verified 2d184cb5

Question Answering for TF trainer (#4320)

jplu committed 5 years ago

Verified ca136186

Fix for #3865. PretrainedTokenizer mapped " do not" into " don't" when .decode(...) is called. Removed the " do not" --> " don't" mapping from clean_up_tokenization(...). (#4024)

lukovnikov committed 5 years ago

Verified 1e51bb71

Newer Older