riaz.somc/transformers
1
0
Fork 0
mirror of https://github.com/huggingface/transformers.git synced 2025-07-31 02:02:21 +06:00
Code Issues Actions 23 Packages Projects Releases Wiki Activity

Return input_ids in ImageGPT feature extractor (#16872)

Browse Source
This commit is contained in:
Sylvain Gugger 2022-04-21 09:09:00 -04:00 committed by GitHub
parent e789418ebe
commit cb555af2c7
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 9 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
src/transformers/models/imagegpt/feature_extraction_imagegpt.py
View File

@ -68,7 +68,7 @@ class ImageGPTFeatureExtractor(FeatureExtractionMixin, ImageFeatureExtractionMix
Whether or not to normalize the input to the range between -1 and +1.
"""
model_input_names = ["pixel_values"]
model_input_names = ["input_ids"]
def __init__(self, clusters, do_resize=True, size=32, resample=Image.BILINEAR, do_normalize=True, **kwargs):
super().__init__(**kwargs)
@ -128,8 +128,7 @@ class ImageGPTFeatureExtractor(FeatureExtractionMixin, ImageFeatureExtractionMix
Returns:
[`BatchFeature`]: A [`BatchFeature`] with the following fields:
- **pixel_values** -- Pixel values to be fed to a model, of shape (batch_size, num_channels, height,
width).
- **input_ids** -- Input IDs to be fed to a model, of shape `(batch_size, height * width)`.
"""
# Input type checking for clearer error
valid_images = False
@ -171,7 +170,7 @@ class ImageGPTFeatureExtractor(FeatureExtractionMixin, ImageFeatureExtractionMix
images = images.reshape(batch_size, -1)
# return as BatchFeature
data = {"pixel_values": images}
data = {"input_ids": images}
encoded_inputs = BatchFeature(data=data, tensor_type=return_tensors)
return encoded_inputs

12
tests/imagegpt/test_feature_extraction_imagegpt.py
View File

@ -161,17 +161,17 @@ class ImageGPTFeatureExtractorIntegrationTest(unittest.TestCase):
# test non-batched
encoding = feature_extractor(images[0], return_tensors="pt")
self.assertIsInstance(encoding.pixel_values, torch.LongTensor)
self.assertEqual(encoding.pixel_values.shape, (1, 1024))
self.assertIsInstance(encoding.input_ids, torch.LongTensor)
self.assertEqual(encoding.input_ids.shape, (1, 1024))
expected_slice = [306, 191, 191]
self.assertEqual(encoding.pixel_values[0, :3].tolist(), expected_slice)
self.assertEqual(encoding.input_ids[0, :3].tolist(), expected_slice)
# test batched
encoding = feature_extractor(images, return_tensors="pt")
self.assertIsInstance(encoding.pixel_values, torch.LongTensor)
self.assertEqual(encoding.pixel_values.shape, (2, 1024))
self.assertIsInstance(encoding.input_ids, torch.LongTensor)
self.assertEqual(encoding.input_ids.shape, (2, 1024))
expected_slice = [303, 13, 13]
self.assertEqual(encoding.pixel_values[1, -3:].tolist(), expected_slice)
self.assertEqual(encoding.input_ids[1, -3:].tolist(), expected_slice)