Skip to content
This repository has been archived by the owner on Dec 16, 2022. It is now read-only.

v2.1.0
Compare
Choose a tag to compare
@dirkgr dirkgr released this 24 Feb 21:20
· 305 commits to main since this release
v2.1.0
8fbc972

What's new

Changed ⚠️

  • coding_scheme parameter is now deprecated in Conll2003DatasetReader, please use convert_to_coding_scheme instead.
  • Support spaCy v3

Added 🎉

  • Added ModelUsage to ModelCard class.
  • Added a way to specify extra parameters to the predictor in an allennlp predict call.
  • Added a way to initialize a Vocabulary from transformers models.
  • Added the ability to use Predictors with multitask models through the new MultiTaskPredictor.
  • Added an example for fields of type ListField[TextField] to apply_token_indexers API docs.
  • Added text_key and label_key parameters to TextClassificationJsonReader class.
  • Added MultiOptimizer, which allows you to use different optimizers for different parts of your model.

Fixed ✅

  • @Registrable.register(...) decorator no longer masks the decorated class's annotations
  • Ensured that MeanAbsoluteError always returns a float metric value instead of a Tensor.
  • Learning rate schedulers that rely on metrics from the validation set were broken in v2.0.0. This
    brings that functionality back.
  • Fixed a bug where the MultiProcessDataLoading would crash when num_workers > 0, start_method = "spawn", max_instances_in_memory not None, and batches_per_epoch not None.
  • Fixed documentation and validation checks for FBetaMultiLabelMetric.
  • Fixed handling of HTTP errors when fetching remote resources with cached_path(). Previously the content would be cached even when
    certain errors - like 404s - occurred. Now an HTTPError will be raised whenever the HTTP response is not OK.
  • Fixed a bug where the MultiTaskDataLoader would crash when num_workers > 0
  • Fixed an import error that happens when PyTorch's distributed framework is unavailable on the system.

Commits

7c6adef Fix worker_info bug when num_workers > 0 (#5013)
9d88f8c Fixes predictors in the multitask case (#4991)
678518a Less opaque registrable annotations (#5010)
4b5fad4 Regex optimizer (#4981)
f091cb9 fix error when torch.distributed not available (#5011)
5974f54 Revert "drop support for Python 3.6 (#5012)" (#5016)
bdb0e20 Update mkdocs-material requirement from <6.3.0,>=5.5.0 to >=5.5.0,<7.1.0 (#5015)
d535de6 Bump mypy from 0.800 to 0.812 (#5007)
099786c Update responses requirement, remove pin on urllib3 (#4783)
b8cfb95 re-write docs commit history on releases (#4968)
c5c9edf Add text_key and label_key to TextClassificationJsonReader (#5005)
a02f67d drop support for Python 3.6 (#5012)
0078c59 Update spaCy to 3.0 (#4953)
be9537f Update CHANGELOG.md
828ee10 Update CHANGELOG.md
1cff6ad update README (#4993)
f8b3807 Add ListField example to apply token indexers (#4987)
7961b8b Ensure mean absolute error metric returns a float (#4983)
da4dba1 raise on HTTP errors in cached_path (#4984)
d4926f5 Inputs to the FBetaMultiLabel metric were copied and pasted wrong (#4975)
d2ae540 Update transformers requirement from <4.3,>=4.1 to >=4.1,<4.4 (#4967)
bf8eeaf Add method to vocab to instantiate from a pretrained transformer (#4958)
9267ce7 Resize transformers word embeddings layer for additional_special_tokens (#4946)
52c23dd Introduce convert_to_coding_scheme and make coding_scheme deprecated in CoNLL2003DatasetReader (#4960)
c418f84 Fixes recording validation metrics for learning rate schedulers that rely on it (#4959)
4535f5c adding ModelUsage, rearranging fields (#4952)
1ace4bb fix bug with MultiProcessDataLoader (#4956)
6f22291 Allows specifying extra arguments for predictors (#4947)
2731db1 tick version for nightly release

Assets 2