NielsRogge/transformers

Pull Requests Commits

More improvements

Niels Rogge committed 2 years ago

74e2bba0

Add first draft

Niels Rogge committed 2 years ago

7cfa7408

Add X-CLIP (#18852)

NielsRogge committed 2 years ago

Verified bb6f6d53

Fix LayoutXLM wrong link in README (#18932)

Devlee247 committed 2 years ago

Verified 9832ac7c

Skip some doctests in quicktour (#18927)

stevhliu committed 2 years ago

Verified 90f6fe91

Add image height and width to ONNX dynamic axes (#18915)

lewtun committed 2 years ago

Verified 6519150c

Starts on a list of external deps required for dev (#18929)

colindean committed 2 years ago

Verified 737f6ad1

Fix XLA fp16 and bf16 error checking (#18913)

ymwangg committed 2 years ago

Verified 63942218

pin TF 2.9.1 for self-hosted CIs (#18925)

ydshieh committed 2 years ago

Verified 6690ba3f

Add DocumentQuestionAnswering pipeline (#18414)

ankrgyl committed 2 years ago

Verified 2ef77421

[DeepSpeed ZeRO3] Fix performance degradation in sharded models (#18911)

tjruwase committed 2 years ago

Verified 3059d80d

remvoe `_create_and_check_torch_fx_tracing` in specific test files (#18667)

ydshieh committed 2 years ago

Verified 10c774cf

TF: final bias as a layer in seq2seq models (replicate TFMarian fix) (#18903)

gante committed 2 years ago

Verified 0eabab09

Update TF fine-tuning docs (#18654)

Rocketknight1 committed 2 years ago

Verified 2b9513fd

update the train_batch_size in case HPO change batch_size_per_device (#18918)

sywangyi committed 2 years ago

Verified d842f2d5

Accelerator end training (#18910)

nbroad1881 committed 2 years ago

Verified 4f299b24

Add checks for more workflow jobs (#18905)

ydshieh committed 2 years ago

Verified 7a811894

[VideoMAE] Improve code examples (#18919)

NielsRogge committed 2 years ago

Verified c25f27fa

Fix incorrect size of input for 1st strided window length in `Perplexity of fixed-length models` (#18906)

ekagra-ranjan committed 2 years ago

Verified 0a632f07

unpin slack_sdk version (#18901)

ydshieh committed 2 years ago

Verified 7d5fde99

Further reduce the number of alls to head for cached objects (#18871)

sgugger committed 2 years ago

Verified 71ff88fa

fixes bugs to handle non-dict output (#18897)

alaradirik committed 2 years ago

Verified 6678350c

Fix `test_tf_encode_plus_sent_to_model` for `LayoutLMv3` (#18898)

ydshieh committed 2 years ago

Verified 998a90bc

Fix decode_input_ids to bare T5Model and improve doc (#18791)

ekagra-ranjan committed 2 years ago

Verified f85acb4d

updating gather function with gather_for_metrics in run_wav2vec2_pretraining (#18877)

arun99481 committed 2 years ago

Verified 3b19c031

Mask t5 relative position bias then head pruned (#17968)

hadaev8 committed 2 years ago

Verified 734b7e2a

Generate: get the correct beam index on eos token (#18851)

gante committed 2 years ago

Verified d4dbd7ca

Update Chinese documentation (#18893)

zkep committed 2 years ago

Verified c6d3daba

Add type hints to XLM-Roberta-XL models (#18475)

asofiaoliveira committed 2 years ago

Verified cfd623a8

Update perf_train_gpu_one.mdx (#18442)

thepurpleowl committed 2 years ago

Verified 17c634fd

Older