* changing find_batch_size to work with tokenizer outputs
trainer_pt_utils.find_batch_size does not recognize the batch size of BatchEncoding objects. This can cause an error when a trainer relies on find_batch_size to report the number of observed examples in the evaluation loop.
* Trigger CI
Co-authored-by: jrenner <joseph.renner@inria.fr>
* Correcting comments to reflect correct tuple order
In order to match the actual order (line 513 and 516, and as accessed in 968), I've changed the order mentioned in comments L962 and L966-967.
* Update modeling_t5.py
Updating another comment as well
* Removing extra space
* Fixing style and quality
* style & quality
* Update src/transformers/models/t5/modeling_t5.py
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* Fix Bart
* Fix Blenderbot{,_small}
* Fix LED
* Fix Marian
* Fix MBart
* Fix Pegasus
* Fix T5
* Add test for generation with head_mask
* Add a common TF test
* Override a test for the LED model as head masking is not yet properly implemented
* Remove all head_masks from input preparation for LED
* Drop masking for T5 as it needs a bit of refactor
The feature extractor does not create tensors on the appropriate device,
so we call `ensure_tensor_on_device` before feeding the processed inputs
to the model.
* [Trainer] Report both steps and num samples per second
* Fix batch number
* Update src/transformers/trainer_utils.py
Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>
* Address review comments
Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>
get_length_grouped_indices() in LengthGroupedSampler and DistributedLengthGroupedSampler
is prohibitively slow for large number of megabatches (in test case takes hours for ~270k
megabatches with 100 items each) due to slow list concatenation with sum(megabatches, []).
Resolves: #11795
Co-authored-by: ctheodoris <cvtheodo@ds.dfci.harvard.edu>
* add separator for windows
* fixes test_is_copy_consistent on Windows
* fixing writing encoding issue on extended test (for Windows)
* resolving comments
Cleaner and more scalable implementation of symbolic tracing with torch.fx, and provides support for new architectures:
- ALBERT
- DistilBERT
- MobileBERT
- MegatronBERT
- GPT2
- GPT Neo
Co-authored-by: Michael Benayoun <michael@huggingface.co>
* Add missing head masking for generate() function
* Add head_mask, decoder_head_mask and cross_attn_head_mask
into prepare_inputs_for_generation for generate() function
for multiple encoder-decoder models.
* Add test_genereate_with_head_masking
* [WIP] Update the new test and handle special cases
* make style
* Omit ProphetNet test so far
* make fix-copies
* [TokenClassification] Label realignment for subword aggregation
Tentative to replace https://github.com/huggingface/transformers/pull/11622/files
- Added `AggregationStrategy`
- `ignore_subwords` and `grouped_entities` arguments are now fused
into `aggregation_strategy`. It makes more sense anyway because
`ignore_subwords=True` with `grouped_entities=False` did not have a
meaning anyway.
- Added 2 new ways to aggregate which are MAX, and AVERAGE
- AVERAGE requires a bit more information than the others, for now this
case is slightly specific, we should keep that in mind for future
changes.
- Testing has been modified to reflect new argument, and to check the
correct deprecation and the new aggregation_strategy.
- Put the testing argument and testing results for aggregation_strategy,
close together, so that readers can understand what is supposed to
happen.
- `aggregate` is now only tested on a small model as it does not mean
anything to test it globally for all models.
- Previous tests are unchanged in desired output.
- Added a new test case that showcases better the difference between the
FIRST, MAX and AVERAGE strategies.
* Wrong framework.
* Addressing three issues.
1- Tags might not follow B-, I- convention, so any tag should work now
(assumed as B-TAG)
2- Fixed an issue with average that leads to a substantial code change.
3- The testing suite was not checking for the "index" key for "none"
strategy. This is now fixed.
The issue is that "O" could not be chosen by AVERAGE strategy because
those tokens were filtered out beforehand, so their relative scores were
not counted in the average. Now filtering on
ignore_labels will happen at the very end of the pipeline fixing
that issue.
It's a bit hard to make sure this stays like that because we do
not have a end-to-end test for that behavior
* Formatting.
* Adding formatting to code + cleaner handling of B-, I- tags.
Co-authored-by: Francesco Rubbo <rubbo.francesco@gmail.com>
Co-authored-by: elk-cloner <rezakakhki.rk@gmail.com>
* Typo.
Co-authored-by: Francesco Rubbo <rubbo.francesco@gmail.com>
Co-authored-by: elk-cloner <rezakakhki.rk@gmail.com>
joerenner