transformers/tests/models/slanext/test_image_processing_slanext.py

# Copyright 2026 The HuggingFace Inc. team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import unittest

import numpy as np

from transformers import is_torch_available, is_vision_available
from transformers.testing_utils import require_torch, require_vision

from ...test_image_processing_common import ImageProcessingTestMixin, prepare_image_inputs


if is_vision_available():
    from PIL import Image


if is_torch_available():
    import torch


class SLANeXtImageProcessingTester:
    def __init__(
        self,
        parent,
        batch_size=7,
        num_channels=3,
        image_size=18,
        min_resolution=10,
        max_resolution=400,
        do_resize=True,
        size=None,
        do_rescale=True,
        rescale_factor=1 / 255,
        do_normalize=True,
        image_mean=[0.485, 0.456, 0.406],
        image_std=[0.229, 0.224, 0.225],
        do_pad=True,
    ):
        size = size if size is not None else {"height": 512, "width": 512}
        self.parent = parent
        self.batch_size = batch_size
        self.num_channels = num_channels
        self.image_size = image_size
        self.min_resolution = min_resolution
        self.max_resolution = max_resolution
        self.do_resize = do_resize
        self.size = size
        self.do_rescale = do_rescale
        self.rescale_factor = rescale_factor
        self.do_normalize = do_normalize
        self.image_mean = image_mean
        self.image_std = image_std
        self.do_pad = do_pad

    def prepare_image_processor_dict(self):
        return {
            "image_mean": self.image_mean,
            "image_std": self.image_std,
            "do_normalize": self.do_normalize,
            "do_resize": self.do_resize,
            "size": self.size,
            "do_pad": self.do_pad,
        }

    def prepare_image_inputs(self, equal_resolution=False, numpify=False, torchify=False):
        return prepare_image_inputs(
            batch_size=self.batch_size,
            num_channels=self.num_channels,
            min_resolution=self.min_resolution,
            max_resolution=self.max_resolution,
            equal_resolution=equal_resolution,
            numpify=numpify,
            torchify=torchify,
        )

    def get_expected_value(self, image_inputs):
        image = image_inputs[0]

        if isinstance(image, Image.Image):
            width, height = image.size
        elif isinstance(image, np.ndarray):
            height, width = image.shape[0], image.shape[1]
        else:
            height, width = image.shape[1], image.shape[2]

        target_size = max(self.size["height"], self.size["width"])
        scale = target_size / max(height, width)
        resize_height = round(height * scale)
        resize_width = round(width * scale)

        if self.do_pad:
            pad_height = max(target_size, resize_height)
            pad_width = max(target_size, resize_width)
            return pad_height, pad_width

        return resize_height, resize_width

    def expected_output_image_shape(self, images):
        height, width = self.get_expected_value(images)
        return self.num_channels, height, width


@require_torch
@require_vision
class SLANeXtImageProcessingTest(ImageProcessingTestMixin, unittest.TestCase):
    def setUp(self):
        super().setUp()
        self.image_processor_tester = SLANeXtImageProcessingTester(self)

    @property
    def image_processor_dict(self):
        return self.image_processor_tester.prepare_image_processor_dict()

    # SLANeXt resizes images adaptively based on aspect ratio, leading to inconsistent output sizes across a batch.
    # Override to skip batched input tests.
    def test_call_pytorch(self):
        for image_processing_class in self.image_processing_classes.values():
            # Initialize image_processing
            image_processing = image_processing_class(**self.image_processor_dict)
            # create random PyTorch tensors
            image_inputs = self.image_processor_tester.prepare_image_inputs(equal_resolution=False, torchify=True)

            for image in image_inputs:
                self.assertIsInstance(image, torch.Tensor)

            # Test not batched input
            encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
            expected_output_image_shape = self.image_processor_tester.expected_output_image_shape([image_inputs[0]])
            self.assertEqual(tuple(encoded_images.shape), (1, *expected_output_image_shape))

    # SLANeXt resizes images adaptively based on aspect ratio, leading to inconsistent output sizes across a batch.
    # Override to skip batched input tests.
    def test_call_numpy(self):
        for image_processing_class in self.image_processing_classes.values():
            # Initialize image_processing
            image_processing = image_processing_class(**self.image_processor_dict)
            # create random numpy tensors
            image_inputs = self.image_processor_tester.prepare_image_inputs(equal_resolution=False, numpify=True)
            for image in image_inputs:
                self.assertIsInstance(image, np.ndarray)

            # Test not batched input
            encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
            expected_output_image_shape = self.image_processor_tester.expected_output_image_shape([image_inputs[0]])
            self.assertEqual(tuple(encoded_images.shape), (1, *expected_output_image_shape))

    # SLANeXt resizes images adaptively based on aspect ratio, leading to inconsistent output sizes across a batch.
    # Override to skip batched input tests.
    def test_call_pil(self):
        for image_processing_class in self.image_processing_classes.values():
            # Initialize image_processing
            image_processing = image_processing_class(**self.image_processor_dict)
            # create random PIL images
            image_inputs = self.image_processor_tester.prepare_image_inputs(equal_resolution=False)
            for image in image_inputs:
                self.assertIsInstance(image, Image.Image)

            # Test not batched input
            encoded_images = image_processing(image_inputs[0], return_tensors="pt").pixel_values
            expected_output_image_shape = self.image_processor_tester.expected_output_image_shape([image_inputs[0]])
            self.assertEqual(tuple(encoded_images.shape), (1, *expected_output_image_shape))

    @unittest.skip(reason="SLANeXtImageProcessorFast does not support 4 channel images yet")
    def test_call_numpy_4_channels(self):
        pass