* implement a simple max_sentencepiece_length into BPE
Add a way for the BPE trainer to behave like the unigram trainer where tokens longer than a certain lenght(default 16 in SPM) to be skipped. this is implemented in unigram trainer but in a different way.
If this code were to be actually integrated some works to be done
Documentation describing the behavior and how it should be set.
Set default==0 so it doesnt act unless set
provide ways in the python binding for the user to set max token length
I was trying to find a way to implement max_sentencepiece_length through pretokenizer split rules and to be honest, its very difficult and regexes can be real slow when operating on the whole training corpus.
* implement a simple max_sentencepiece_length into BPE
Add a way for the BPE trainer to behave like the unigram trainer where tokens longer than a certain lenght(default 16 in SPM) to be skipped. this is implemented in unigram trainer but in a different way.
If this code were to be actually integrated some works to be done
Documentation describing the behavior and how it should be set.
Set default==0 so it doesnt act unless set
provide ways in the python binding for the user to set max token length
I was trying to find a way to implement max_sentencepiece_length through pretokenizer split rules and to be honest, its very difficult and regexes can be real slow when operating on the whole training corpus.
* utilize Option<u16> for safer code.
* Other version.
* Update trainer.rs
clarify with type usize propagate max_length option
* change max_length into more descriptive name
in the documentation
https://huggingface.co/docs/tokenizers/api/trainers
unigramtrainer uses max_piece_length for similar function.
since BPE the underlying concept is merges, using max_merge_length as the variable name could prove more descriptive.
* change variable name in trainer.rs
change max_merge_length into max_token_length
* Update trainer.rs
add several max_token_length declaration that were missing.
impl BpeTrainerBuilder
struct BpeTrainer
Add explanation for variable shadowing.
* Update trainer.rs
Move default definition of max_token_length to proper location. adjust downstream variable initializations accordingly.
* add max_token_length test
* Add bpe direct assert test
* Update trainer.rs
clarified test documentation
* Creating the bindings.
* Fix the default.
* Re-adding missing package-lock which I accidentally removed.
* ..
* Fixing trainer test.
* Fix.
---------
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
* Adding 2 new decoders:
- Fuse will simply concatenate all tokens into 1 string
- Strip will remove n char from left or right
Sequence(Replace("_", " "), Fuse(), Strip(1, 0)) should be what we want
for the `Metaspace` thing.
- Note: Added a new dependency from better parsing of decoders.
This is due to untagged enums which can match anything the `MustBe`
ensure there's no issue between Fuse and ByteFallback.
Since both are new the chances for backward incompatibility is low.
* Fixing picking/unpickling (using default args.).
* Stub.
* Black.
* Fixing node.
* Adding ByteFallback support for `tokenizers`.
Two items added:
- A flag `byte_fallback` for the `BPE` model. This will be in charge
of using `<0x61>` instead of unk on unknown tokens.
- A ByteFallback decoder, which will be in charge of putting everything
back into string whenever possible. Showing � when the byte decoding
fails (behavior checked against LlamaTokenizer in `transformers`.
* Update rustdoc.
* Clippy + Add BPE(byte_fallback) into bindings.
* Stupid file.
* Test artifacts removed.
* Update stub.
* Fix.
* Bad file.
* CRITICAL FIX: wrapper order because of untagged....
* Remove prints.
* Fixing <16 byte fallback.
* [fix] Use unk_token
In SentencePieceBPETokenizer, when Vocab or merges is None, unk_token cannot be used.
* [fix] If unk_token is None, this case is also considered.
* Update bindings/python/py_src/tokenizers/implementations/sentencepiece_bpe.py
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
* [FIX] In CharBPETokenizer, Use unk_token.
In CharBPETokenizer, when Vocab or merges is None, unk_token cannot be used.
* Update bindings/python/py_src/tokenizers/implementations/char_level_bpe.py
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
* Update bindings/python/py_src/tokenizers/implementations/char_level_bpe.py
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
* [fix] Use unk_token
In SentencePieceBPETokenizer, when Vocab or merges is None, unk_token cannot be used.
* [fix] If unk_token is None, this case is also considered.
* Update bindings/python/py_src/tokenizers/implementations/sentencepiece_bpe.py
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
* Adding rust audit.
* Update clap version + derive_builder (they clashed).
* Ignoring specific CVE which can be ignored
https://github.com/Azure/iot-identity-service/issues/481
* Updating python lock.
* Revert `derive-builder` update.
* Adding back help msg.
* New version.
The actual release will happen *before* PyO3 0.17.2 because
the tests were ran before than.
* Manylinux2014 necessary now with Rust 1.64.
* Fixing roberta type ids (everything is zero).
* We need to fix type_ids for all sequence even when not changing
anything else.
* Fixing tests hopefully better.
* Removing dead file.
* Checking that we can distribute with static python embedding for
manylinux
* Many linux embed interpreter.
* Building wheels manylinux with static embedding
* Better script.
* typo.
* Using a dummy feature?
* default features ?
* Back into order.
* Fixing manylinux ??.
* Local dir.
* Missing star.
* Makedir ?
* Monkey coding this.
* extension module ?
* Building with default features `RustExtension`.
* bdist_wheel + rustextension any better ?
* update rust-py version.
* Forcing extension module.
* No default features.
* Remove py37 out of spite
* Revert "Remove py37 out of spite"
This reverts commit 6ab7facd792b59c2e30be82fe42816d24c32cf0d.
* Really extraneous feature.
* Fix build wheels.
* Putting things back in place.
* Upgrade pyo3 to 0.15
Rebase-conflicts-fixed-by: H. Vetinari <h.vetinari@gmx.com>
* Upgrade pyo3 to 0.16
Rebase-conflicts-fixed-by: H. Vetinari <h.vetinari@gmx.com>
* Install Python before running cargo clippy
* Fix clippy warnings
* Use `PyArray_Check` instead of downcasting to `PyArray1<u8>`
* Enable `auto-initialize` of pyo3 to fix `cargo test
--no-default-features`
* Fix some test cases
Why do they change?
* Refactor and add SAFETY comments to `PyArrayUnicode`
Replace deprecated `PyUnicode_FromUnicode` with `PyUnicode_FromKindAndData`
Co-authored-by: messense <messense@icloud.com>
* Fixing the vocab size of the trained Unigram model
* add test for the vocab size of the trained Unigram model
* Revert "add test for the vocab size of the trained Unigram model"
This reverts commit fb8955c831b357d1037548ceaa8789734d544646.
* Fixing the vocab size of the trained Unigram model
* format codes
* get the position of vocab-size calculation out of loop
* TMP.
* Adding support for pickling Python trainers.
* Remove not warranted files + missed naming updates.
* Stubbing.
* Making sure serialized format is written in python tests.
