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

added deit onnx config #16887

Merged
merged 5 commits into from
Apr 25, 2022
Merged

added deit onnx config #16887

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
1ad92cb
added deit onnx config
0xrushi Apr 22, 2022
d914ad6
tests fix
0xrushi Apr 22, 2022
695696c
masked modelling
0xrushi Apr 22, 2022
39ed9ee
indentation fix
0xrushi Apr 22, 2022
bb4de21
removed datasets
0xrushi Apr 25, 2022
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
1 change: 1 addition & 0 deletions datasets
Open in desktop
Submodule datasets added at 4134e8
1 change: 1 addition & 0 deletions docs/source/en/serialization.mdx
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,6 +55,7 @@ Ready-made configurations include the following architectures:
- CamemBERT
- Data2VecText
- Data2VecVision
- DeiT
- DistilBERT
- ELECTRA
- FlauBERT
Expand Down
4 changes: 2 additions & 2 deletions src/transformers/models/deit/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,7 @@


_import_structure = {
"configuration_deit": ["DEIT_PRETRAINED_CONFIG_ARCHIVE_MAP", "DeiTConfig"],
"configuration_deit": ["DEIT_PRETRAINED_CONFIG_ARCHIVE_MAP", "DeiTConfig", "DeiTOnnxConfig"],
}

if is_vision_available():
Expand All @@ -39,7 +39,7 @@


if TYPE_CHECKING:
from .configuration_deit import DEIT_PRETRAINED_CONFIG_ARCHIVE_MAP, DeiTConfig
from .configuration_deit import DEIT_PRETRAINED_CONFIG_ARCHIVE_MAP, DeiTConfig, DeiTOnnxConfig

if is_vision_available():
from .feature_extraction_deit import DeiTFeatureExtractor
Expand Down
22 changes: 22 additions & 0 deletions src/transformers/models/deit/configuration_deit.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,13 @@
# limitations under the License.
""" DeiT model configuration"""

from collections import OrderedDict
from typing import Mapping

from packaging import version

from ...configuration_utils import PretrainedConfig
from ...onnx import OnnxConfig
from ...utils import logging


Expand Down Expand Up @@ -120,3 +126,19 @@ def __init__(
self.num_channels = num_channels
self.qkv_bias = qkv_bias
self.encoder_stride = encoder_stride


class DeiTOnnxConfig(OnnxConfig):
torch_onnx_minimum_version = version.parse("1.11")

@property
def inputs(self) -> Mapping[str, Mapping[int, str]]:
return OrderedDict(
[
("pixel_values", {0: "batch", 1: "sequence"}),
]
)

@property
def atol_for_validation(self) -> float:
return 1e-4
2 changes: 2 additions & 0 deletions src/transformers/onnx/features.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@
from ..models.blenderbot_small import BlenderbotSmallOnnxConfig
from ..models.camembert import CamembertOnnxConfig
from ..models.data2vec import Data2VecTextOnnxConfig
from ..models.deit import DeiTOnnxConfig
from ..models.distilbert import DistilBertOnnxConfig
from ..models.electra import ElectraOnnxConfig
from ..models.flaubert import FlaubertOnnxConfig
Expand Down Expand Up @@ -286,6 +287,7 @@ class FeaturesManager:
),
"vit": supported_features_mapping("default", "image-classification", onnx_config_cls=ViTOnnxConfig),
"beit": supported_features_mapping("default", "image-classification", onnx_config_cls=BeitOnnxConfig),
"deit": supported_features_mapping("default", "image-classification", onnx_config_cls=DeiTOnnxConfig),
Copy link
Contributor

Choose a reason for hiding this comment

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

I saw that DeiT also have a DeiTForMaskedImageModeling implementation, is it a feature also available in onnx ?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

BeiT also has BeiTForMaskedImageModeling, but there isn't anything for that in features.py

Copy link
Member

Choose a reason for hiding this comment

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

You can add something like

"masked-im": AutoModelForMaskedImageModeling

to the _TASKS_TO_AUTOMODELS attribute of the FeaturesManager. I don't think we currently support this feature for TensorFlow, so no need to include this in __TASKS_TO_TF_AUTOMODELS.

Feel free to then include this feature as part of vit, deit, and beit

"blenderbot": supported_features_mapping(
"default",
"default-with-past",
Expand Down
1 change: 1 addition & 0 deletions tests/onnx/test_onnx_v2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -181,6 +181,7 @@ def test_values_override(self):
("xlm-roberta", "xlm-roberta-base"),
("layoutlm", "microsoft/layoutlm-base-uncased"),
("vit", "google/vit-base-patch16-224"),
("deit", "facebook/deit-small-patch16-224"),
("beit", "microsoft/beit-base-patch16-224"),
("data2vec-text", "facebook/data2vec-text-base"),
}
Expand Down