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

Convert tokenizers with openvino_tokenizers #500

Closed
wants to merge 2 commits into from
+27 −0
Closed

Convert tokenizers with openvino_tokenizers #500

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
feba5bf
Convert tokenizers with openvino_tokenizers
slyalin Jan 5, 2024
4e7bfa9
Update optimum/exporters/openvino/__main__.py
slyalin Jan 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
27 changes: 27 additions & 0 deletions optimum/exporters/openvino/__main__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,7 @@

from requests.exceptions import ConnectionError as RequestsConnectionError
from transformers import AutoConfig, AutoTokenizer
from openvino import save_model

from optimum.exporters import TasksManager
from optimum.exporters.onnx import __main__ as optimum_main
Expand Down Expand Up @@ -46,6 +47,24 @@
logger = logging.getLogger(__name__)


def tokenizer_export(
Copy link
Collaborator

Choose a reason for hiding this comment

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

I think it should be in convert.py file together with other conversion functions.

Copy link
Contributor

Choose a reason for hiding this comment

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

Moved.

tokenizer,
output: Union[str, Path],
suffix: Optional[str] = ""
):
try:
from openvino_tokenizers import convert_tokenizer
ov_tokenizer, ov_detokenizer = convert_tokenizer(tokenizer, with_detokenizer=True)
if isinstance(output, str):
output = Path(output)
tokenizer_path = output.joinpath("openvino_tokenizer" + suffix + ".xml")
detokenizer_path = output.joinpath("openvino_detokenizer" + suffix + ".xml")
save_model(ov_tokenizer, tokenizer_path)
save_model(ov_detokenizer, detokenizer_path)
except Exception as exception:
print("[ WARNING ] OpenVINO tokenizer/detokenizer models couldn't be exported because of exception:", exception)
Copy link
Collaborator

Choose a reason for hiding this comment

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

Suggested change
print("[ WARNING ] OpenVINO tokenizer/detokenizer models couldn't be exported because of exception:", exception)
logger.warning(OpenVINO tokenizer/detokenizer models couldn't be exported because of exception:", exception)

Copy link
Contributor

Choose a reason for hiding this comment

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

Fixed.



def main_export(
model_name_or_path: str,
output: Union[str, Path],
Expand Down Expand Up @@ -328,6 +347,12 @@ class StoreAttr(object):
if generation_config is not None:
generation_config.save_pretrained(output)
maybe_save_preprocessors(model_name_or_path, output)
try:
# TODO: Avoid loading the tokenizer again if loaded before
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
tokenizer_export(tokenizer, output)
except:
print("[ WARNING ] Could not load tokenizer using specified model ID or path. OpenVINO tokenizer/detokenizer models won't be generated.")
Comment on lines +350 to +355
Copy link
Collaborator

Choose a reason for hiding this comment

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

tokenizer already loaded in maybe_load_preprocessors function, I recommend to check this function result instead. Also possibly you should take into account trust_remote_code parameter if you want load tokenizer explicitly

Copy link
Contributor

@apaniukov apaniukov Jan 10, 2024 •
edited
Loading

Choose a reason for hiding this comment

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

Reuse tokenizer from maybe_load_preprocessors result. But a new problem is that there are two identical tokenizers - one from AutoTokenizer and one from AutoProcessor. I haven't figured out how to deduplicate them yet.

Copy link
Contributor

Choose a reason for hiding this comment

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

Here is the PR: slyalin#2


if model.config.is_encoder_decoder and task.startswith("text-generation"):
raise ValueError(
Expand Down Expand Up @@ -358,10 +383,12 @@ class StoreAttr(object):
tokenizer = getattr(model, "tokenizer", None)
if tokenizer is not None:
tokenizer.save_pretrained(output.joinpath("tokenizer"))
tokenizer_export(tokenizer, output)

tokenizer_2 = getattr(model, "tokenizer_2", None)
if tokenizer_2 is not None:
tokenizer_2.save_pretrained(output.joinpath("tokenizer_2"))
tokenizer_export(tokenizer, output, "_2")

model.save_config(output)

Expand Down
Loading