Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

add truncation in text translator #472

Merged
merged 4 commits into from
May 3, 2024
Merged

add truncation in text translator #472

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
64cb547
add truncation in text translator
avidale Apr 24, 2024
f9804f7
add a test case for truncation
avidale Apr 26, 2024
3651108
Merge branch 'main' into truncate-in-translator
avidale Apr 26, 2024
e980dbc
rename the parameter
avidale Apr 26, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
11 changes: 11 additions & 0 deletions src/fairseq2/generation/text.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -153,13 +153,15 @@ class TextTranslator:
_converter: SequenceToTextConverter
_pad_idx: int
_source_text_encoder: TextTokenEncoder
max_src_len: Optional[int]
avidale marked this conversation as resolved.
Show resolved Hide resolved

def __init__(
self,
generator: Seq2SeqGenerator,
tokenizer: TextTokenizer,
source_lang: Optional[str] = None,
target_lang: Optional[str] = None,
max_src_len: Optional[int] = None,
avidale marked this conversation as resolved.
Show resolved Hide resolved
) -> None:
"""
:param generator:
Expand All @@ -170,6 +172,8 @@ def __init__(
The source language.
:param target_lang:
The target language.
:param max_src_len:
The number of tokens above which the source sequence gets truncated (None for no truncation)
"""
self._converter = SequenceToTextConverter(
generator, tokenizer, "translation", target_lang
Expand All @@ -188,6 +192,7 @@ def __init__(
self._source_text_encoder = tokenizer.create_encoder(
task="translation", lang=source_lang, mode="source", device=device
)
self.max_src_len = max_src_len

def __call__(self, source_text: str) -> Tuple[str, Seq2SeqGeneratorOutput]:
"""
Expand All @@ -200,6 +205,9 @@ def __call__(self, source_text: str) -> Tuple[str, Seq2SeqGeneratorOutput]:
"""
source_seq = self._source_text_encoder(source_text)

if self.max_src_len and source_seq.shape[0] > self.max_src_len:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Although it would be a very edge case; if max_src_len is 0, then the first condition if self.max_src_len will be false, but maybe the user deliberately wants to trim to zero length (although does not make much sense, still expected since our parameter contract says that this parameter is ignored only if it is None). The "right" check would be if self.max_src_len is not None and ...

Copy link
Contributor Author

@avidale avidale Apr 26, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If the source is trimmed to 0 elements, the translation is going to fail, because the encoder would consume an empty tensor. Therefore, I don't want to apply truncation to zero, and instead will change the parameter contract to explicitly truncate only to a positive number of tokens.

source_seq = source_seq[: self.max_src_len]

return self._converter(source_seq)

def batch_translate(
Expand All @@ -220,6 +228,9 @@ def batch_translate(

source_seq_list = [self._source_text_encoder(t) for t in source_texts]

if self.max_src_len:
source_seq_list = [seq[: self.max_src_len] for seq in source_seq_list]

source_seqs, source_padding_mask = pad_seqs(source_seq_list, self._pad_idx)

return self._converter.batch_convert(source_seqs, source_padding_mask)
Expand Down
18 changes: 17 additions & 1 deletion tests/integration/models/test_nllb.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,6 +6,7 @@

from typing import Final

import pytest
import torch

from fairseq2.generation import BeamSearchSeq2SeqGenerator, TextTranslator
Expand All @@ -25,7 +26,7 @@ def test_load_dense_distill_600m() -> None:

tokenizer = load_nllb_tokenizer(model_name, progress=False)

generator = BeamSearchSeq2SeqGenerator(model, echo_prompt=True)
generator = BeamSearchSeq2SeqGenerator(model, echo_prompt=True, max_seq_len=128)

translator = TextTranslator(
generator, tokenizer, source_lang="eng_Latn", target_lang="deu_Latn"
Expand All @@ -34,3 +35,18 @@ def test_load_dense_distill_600m() -> None:
text, _ = translator(ENG_SENTENCE)

assert text == DEU_SENTENCE

# testing that truncation prevents length-related errors
with pytest.raises(
ValueError, match="The input sequence length must be less than or equal"
):
text, _ = translator(ENG_SENTENCE * 20)

translator = TextTranslator(
generator,
tokenizer,
source_lang="eng_Latn",
target_lang="deu_Latn",
max_src_len=1024,
)
text, _ = translator(ENG_SENTENCE * 20)
Loading