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

fix pylist #1673

Merged
merged 7 commits into from
Nov 5, 2024
Merged

fix pylist #1673

Changes from 4 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
d50f9ae
fix pylist
ArthurZucker Nov 5, 2024
c2dc6b0
add comment about why we use PySequence
ArthurZucker Nov 5, 2024
495d430
style
ArthurZucker Nov 5, 2024
8c96f47
fix encode batch fast as well
ArthurZucker Nov 5, 2024
e1169ee
Update bindings/python/src/tokenizer.rs
ArthurZucker Nov 5, 2024
6056e0c
fix with capacity
ArthurZucker Nov 5, 2024
2da036e
stub :)
ArthurZucker Nov 5, 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
54 changes: 27 additions & 27 deletions bindings/python/src/tokenizer.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -995,7 +995,9 @@ impl PyTokenizer {
}

/// Encode the given batch of inputs. This method accept both raw text sequences
/// as well as already pre-tokenized sequences.
/// as well as already pre-tokenized sequences. The reason we use `PySequence` is
/// because it allows type checking with zero-cost (according to PyO3) as we don't
/// have to convert to check.
///
/// Example:
/// Here are some examples of the inputs that are accepted::
Expand Down Expand Up @@ -1030,25 +1032,24 @@ impl PyTokenizer {
fn encode_batch(
&self,
py: Python<'_>,
input: Bound<'_, PyList>,
input: Bound<'_, PySequence>,
is_pretokenized: bool,
add_special_tokens: bool,
) -> PyResult<Vec<PyEncoding>> {
let input: Vec<tk::EncodeInput> = input
.into_iter()
.map(|o| {
let input: tk::EncodeInput = if is_pretokenized {
o.extract::<PreTokenizedEncodeInput>()?.into()
} else {
o.extract::<TextEncodeInput>()?.into()
};
Ok(input)
})
.collect::<PyResult<Vec<tk::EncodeInput>>>()?;
let mut items = Vec::<tk::EncodeInput>::new();
ArthurZucker marked this conversation as resolved.
Show resolved Hide resolved
for i in 0..input.len()? {
let item = input.get_item(i)?;
let item: tk::EncodeInput = if is_pretokenized {
item.extract::<PreTokenizedEncodeInput>()?.into()
} else {
item.extract::<TextEncodeInput>()?.into()
};
items.push(item);
}
py.allow_threads(|| {
ToPyResult(
self.tokenizer
.encode_batch_char_offsets(input, add_special_tokens)
.encode_batch_char_offsets(items, add_special_tokens)
.map(|encodings| encodings.into_iter().map(|e| e.into()).collect()),
)
.into()
Expand Down Expand Up @@ -1091,25 +1092,24 @@ impl PyTokenizer {
fn encode_batch_fast(
&self,
py: Python<'_>,
input: Bound<'_, PyList>,
input: Bound<'_, PySequence>,
is_pretokenized: bool,
add_special_tokens: bool,
) -> PyResult<Vec<PyEncoding>> {
let input: Vec<tk::EncodeInput> = input
.into_iter()
.map(|o| {
let input: tk::EncodeInput = if is_pretokenized {
o.extract::<PreTokenizedEncodeInput>()?.into()
} else {
o.extract::<TextEncodeInput>()?.into()
};
Ok(input)
})
.collect::<PyResult<Vec<tk::EncodeInput>>>()?;
let mut items = Vec::<tk::EncodeInput>::new();
for i in 0..input.len()? {
let item = input.get_item(i)?;
let item: tk::EncodeInput = if is_pretokenized {
item.extract::<PreTokenizedEncodeInput>()?.into()
} else {
item.extract::<TextEncodeInput>()?.into()
};
items.push(item);
}
py.allow_threads(|| {
ToPyResult(
self.tokenizer
.encode_batch_fast(input, add_special_tokens)
.encode_batch_fast(items, add_special_tokens)
.map(|encodings| encodings.into_iter().map(|e| e.into()).collect()),
)
.into()
Expand Down
Loading