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 do_convert_rgb to vit #34523

Merged
merged 8 commits into from
Nov 19, 2024
Merged

Add do_convert_rgb to vit #34523

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
0535d7c
Add: do_convert_rgb
jp1924 Oct 31, 2024
052a89c
Add: doc string
jp1924 Oct 31, 2024
1600d26
Update src/transformers/models/vit/image_processing_vit.py
jp1924 Nov 5, 2024
71c7fe2
Update src/transformers/models/vit/image_processing_vit.py
jp1924 Nov 5, 2024
3fc89df
Update src/transformers/models/vit/image_processing_vit.py
jp1924 Nov 5, 2024
a922a4e
Add: do_convert_rgb to fast
jp1924 Nov 5, 2024
e9503ec
Merge branch 'main' into add_do_rgb_convert_to_vit
jp1924 Nov 8, 2024
ae573cf
Add: convert_to_rgb
jp1924 Nov 12, 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
13 changes: 12 additions & 1 deletion src/transformers/models/vit/image_processing_vit.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,7 +19,7 @@
import numpy as np

from ...image_processing_utils import BaseImageProcessor, BatchFeature, get_size_dict
from ...image_transforms import resize, to_channel_dimension_format
from ...image_transforms import convert_to_rgb, resize, to_channel_dimension_format
from ...image_utils import (
IMAGENET_STANDARD_MEAN,
IMAGENET_STANDARD_STD,
Expand Down Expand Up @@ -68,6 +68,8 @@ class ViTImageProcessor(BaseImageProcessor):
image_std (`float` or `List[float]`, *optional*, defaults to `IMAGENET_STANDARD_STD`):
Standard deviation to use if normalizing the image. This is a float or list of floats the length of the
number of channels in the image. Can be overridden by the `image_std` parameter in the `preprocess` method.
do_convert_rgb (`bool`, *optional*):
Whether to convert the image to RGB.
"""

model_input_names = ["pixel_values"]
Expand All @@ -82,6 +84,7 @@ def __init__(
do_normalize: bool = True,
image_mean: Optional[Union[float, List[float]]] = None,
image_std: Optional[Union[float, List[float]]] = None,
do_convert_rgb: Optional[bool] = None,
**kwargs,
) -> None:
super().__init__(**kwargs)
Expand All @@ -95,6 +98,7 @@ def __init__(
self.rescale_factor = rescale_factor
self.image_mean = image_mean if image_mean is not None else IMAGENET_STANDARD_MEAN
self.image_std = image_std if image_std is not None else IMAGENET_STANDARD_STD
self.do_convert_rgb = do_convert_rgb

def resize(
self,
Expand Down Expand Up @@ -159,6 +163,7 @@ def preprocess(
return_tensors: Optional[Union[str, TensorType]] = None,
data_format: Union[str, ChannelDimension] = ChannelDimension.FIRST,
input_data_format: Optional[Union[str, ChannelDimension]] = None,
do_convert_rgb: Optional[bool] = None,
):
"""
Preprocess an image or batch of images.
Expand Down Expand Up @@ -203,6 +208,8 @@ def preprocess(
- `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
- `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
- `"none"` or `ChannelDimension.NONE`: image in (height, width) format.
do_convert_rgb (`bool`, *optional*, defaults to `self.do_convert_rgb`):
Whether to convert the image to RGB.
"""
do_resize = do_resize if do_resize is not None else self.do_resize
do_rescale = do_rescale if do_rescale is not None else self.do_rescale
Expand All @@ -211,6 +218,7 @@ def preprocess(
rescale_factor = rescale_factor if rescale_factor is not None else self.rescale_factor
image_mean = image_mean if image_mean is not None else self.image_mean
image_std = image_std if image_std is not None else self.image_std
do_convert_rgb = do_convert_rgb if do_convert_rgb is not None else self.do_convert_rgb

size = size if size is not None else self.size
size_dict = get_size_dict(size)
Expand All @@ -233,6 +241,9 @@ def preprocess(
resample=resample,
)

if do_convert_rgb:
images = [convert_to_rgb(image) for image in images]

# All transformations expect numpy arrays.
images = [to_numpy_array(image) for image in images]

Expand Down
13 changes: 12 additions & 1 deletion src/transformers/models/vit/image_processing_vit_fast.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,7 +20,7 @@
from ...image_processing_base import BatchFeature
from ...image_processing_utils import get_size_dict
from ...image_processing_utils_fast import BaseImageProcessorFast, SizeDict
from ...image_transforms import FusedRescaleNormalize, NumpyToTensor, Rescale
from ...image_transforms import FusedRescaleNormalize, NumpyToTensor, Rescale, convert_to_rgb
from ...image_utils import (
IMAGENET_STANDARD_MEAN,
IMAGENET_STANDARD_STD,
Expand Down Expand Up @@ -76,6 +76,8 @@ class ViTImageProcessorFast(BaseImageProcessorFast):
image_std (`float` or `List[float]`, *optional*, defaults to `IMAGENET_STANDARD_STD`):
Standard deviation to use if normalizing the image. This is a float or list of floats the length of the
number of channels in the image. Can be overridden by the `image_std` parameter in the `preprocess` method.
do_convert_rgb (`bool`, *optional*):
Whether to convert the image to RGB.
"""

model_input_names = ["pixel_values"]
Expand All @@ -101,6 +103,7 @@ def __init__(
do_normalize: bool = True,
image_mean: Optional[Union[float, List[float]]] = None,
image_std: Optional[Union[float, List[float]]] = None,
do_convert_rgb: Optional[bool] = None,
**kwargs,
) -> None:
super().__init__(**kwargs)
Expand All @@ -114,6 +117,7 @@ def __init__(
self.rescale_factor = rescale_factor
self.image_mean = image_mean if image_mean is not None else IMAGENET_STANDARD_MEAN
self.image_std = image_std if image_std is not None else IMAGENET_STANDARD_STD
self.do_convert_rgb = do_convert_rgb

def _build_transforms(
self,
Expand Down Expand Up @@ -199,6 +203,7 @@ def preprocess(
return_tensors: Optional[Union[str, TensorType]] = "pt",
data_format: Union[str, ChannelDimension] = ChannelDimension.FIRST,
input_data_format: Optional[Union[str, ChannelDimension]] = None,
do_convert_rgb: Optional[bool] = None,
**kwargs,
):
"""
Expand Down Expand Up @@ -237,6 +242,8 @@ def preprocess(
- `"channels_first"` or `ChannelDimension.FIRST`: image in (num_channels, height, width) format.
- `"channels_last"` or `ChannelDimension.LAST`: image in (height, width, num_channels) format.
- `"none"` or `ChannelDimension.NONE`: image in (height, width) format.
do_convert_rgb (`bool`, *optional*):
Whether to convert the image to RGB.
"""
do_resize = do_resize if do_resize is not None else self.do_resize
do_rescale = do_rescale if do_rescale is not None else self.do_rescale
Expand All @@ -246,6 +253,7 @@ def preprocess(
image_mean = image_mean if image_mean is not None else self.image_mean
image_std = image_std if image_std is not None else self.image_std
size = size if size is not None else self.size
do_convert_rgb = do_convert_rgb if do_convert_rgb is not None else self.do_convert_rgb
# Make hashable for cache
size = SizeDict(**size)
image_mean = tuple(image_mean) if isinstance(image_mean, list) else image_mean
Expand All @@ -271,6 +279,9 @@ def preprocess(
image_type=image_type,
)

if do_convert_rgb:
images = [convert_to_rgb(image) for image in images]

transforms = self.get_transforms(
do_resize=do_resize,
do_rescale=do_rescale,
Expand Down
Loading