* Switch to plain unittest for skipping slow tests.
Add a RUN_SLOW environment variable for running them.
* Switch to plain unittest for PyTorch dependency.
* Switch to plain unittest for TensorFlow dependency.
* Avoid leaking open files in the test suite.
This prevents spurious warnings when running tests.
* Fix unicode warning on Python 2 when running tests.
The warning was:
UnicodeWarning: Unicode equal comparison failed to convert both arguments to Unicode - interpreting them as being unequal
* Support running PyTorch tests on a GPU.
Reverts 27e015bd.
* Tests no longer require pytest.
* Make tests pass on cuda
Whenever target_mapping is provided to the input, XLNet outputs two different attention streams.
Based on that the attention output would be on of the two:
- a list of tensors (usual case for most transformers)
- a list of 2-tuples of tensors, one tesor for each of attention streams
Docs and unit-tests have been updated
Custom schedulers are currently initiated by wrapping Pytorch's LambdaLR
class and passing a method of the wrapping class to the __init__
function of LambdaLR. This approach is not appropriate for several
reasons:
1. one does not need to define a class when it only defines a
__init__() method;
2. instantiating the parent class by passing a method of the child class
creates a cyclical reference which leads to memory leaks. See issues #1742 and #1134.
In this commit we replace the wrapper classes with functions that
instantiate `LambdaLR` with a custom learning rate function. We use a
closure to specify the parameter of the latter. We also do a bit of
renaming within the function to explicit the behaviour and removed
docstrings that were subsequently not necessary.
As pointed out in #1545, when using an uncased model, and adding
a new uncased token, the tokenizer does not correctly identify this
in the case that the input text contains the token in a cased format.
For instance, if we load bert-base-uncased into BertTokenizer, and
then use .add_tokens() to add "cool-token", we get the expected
result for .tokenize('this is a cool-token'). However, we get a
possibly unexpected result for .tokenize('this is a cOOl-Token'),
which in fact mirrors the result for the former from before the new
token was added.
This commit adds
- functionality to PreTrainedTokenizer to handle this
situation in case a tokenizer (currently Bert, DistilBert,
and XLNet) has the do_lower_case=True kwarg by:
1) lowercasing tokens added with .add_tokens()
2) lowercasing text at the beginning of .tokenize()
- new common test case for tokenizers
https://github.com/huggingface/transformers/issues/1545
