Add normalize to image transforms module #19544
amyeroberts
merged
219 commits into
huggingface:main
from
amyeroberts:image-transforms-add-normalize
Adapt FE methods to transforms library
a94c5377
Mixin for saving the image processor
932f291f
Base processor skeleton
54aed8b3
BatchFeature for packaging image processor outputs
ba55c899
Initial image processor for GLPN
4b430d4d
REmove accidental import
b1c8b59f
Fixup and docs
daf069a2
Mixin for saving the image processor
95b4a6af
Fixup and docs
6f7ef562
Import BatchFeature from feature_extraction_utils
b9ce4a00
Merge branch 'image-processor-mixin' of github.com:amyeroberts/transf…
f02ae6a7
Fixup and docs
6b678fb4
Fixup and docs
db934377
Fixup and docs
bd890d59
Fixup and docs
4b27a340
BatchFeature for packaging image processor outputs
ff0d49ec
Import BatchFeature from feature_extraction_utils
2c2fa9aa
Merge branch 'image-processor-mixin' into base-image-processor-class
b9f78379
Resolve conflicts
346270d0
Import BatchFeature from feature_extraction_utils
7faf2e69
Fixup and docs
ccc15fb6
Fixup and docs
c8f8eb6e
BatchFeature for packaging image processor outputs
90093f4b
Import BatchFeature from feature_extraction_utils
d89c0513
Fixup and docs
9bc91578
Mixin for saving the image processor
6ec382ac
Fixup and docs
56ee6ad2
Merge branch 'image-batch-feature' into image-processor-glpn
38ebb509
Add rescale back and remove ImageType
6b88d5f8
fix import mistake
67077f1d
Fix enum var reference
82712c76
Merge branch 'image-transforms-library' into image-processor-mixin
71d666d9
Merge branch 'image-processor-mixin' into base-image-processor-class
fb6438c9
Merge branch 'base-image-processor-class' into image-batch-feature
ffe71b61
Merge branch 'image-batch-feature' into image-processor-glpn
cc480e82
Can transform and specify image data format
b997a984
Merge branch 'image-transforms-library' into image-processor-mixin
91064434
Remove redundant function
1b3cf656
Update reference
28604600
Merge branch 'image-transforms-library' into image-processor-mixin
3e1077bd
Merge branch 'image-processor-mixin' into base-image-processor-class
4264d1a1
Merge in branch and remove conflicts
fb5dcd6d
Add in rescaling
43f561d0
Data format flag for rescale
60c56e5d
Fix typo
9294dbce
Fix dimension check
654cf932
Merge branch 'image-transforms-library' into image-processor-mixin
13607324
Merge branch 'image-processor-mixin' into base-image-processor-class
936de657
Merge branch 'base-image-processor-class' into image-batch-feature
627c0482
Merge branch 'image-batch-feature' into image-processor-glpn
1b64c805
Fixes to make IP and FE outputs match
88b82e93
Add tests for transforms
3ea27aa9
Add test for utils
84fdd074
Merge branch 'image-transforms-library' into image-processor-mixin
10d56b11
Update some docstrings
392e980d
Merge branch 'image-processor-mixin' into base-image-processor-class
2117b94b
Resole merge conflicts
68de9525
Merge branch 'image-batch-feature' into image-processor-glpn
5208680a
Make sure in channels last before converting to PIL
a28ac88f
Merge branch 'image-transforms-library' into image-processor-mixin
2ead9e5e
Merge branch 'image-processor-mixin' into base-image-processor-class
9514d54c
Merge branch 'base-image-processor-class' into image-batch-feature
8f63b764
Merge branch 'image-batch-feature' into image-processor-glpn
46a9c743
Remove default to numpy batching
082e4ff9
Fix up
bf733582
Add docstring and model_input_types
34b6b2fa
Use feature processor config from hub
71502937
Merge branch 'image-processor-mixin' into base-image-processor-class
8678c13a
Merge branch 'base-image-processor-class' into image-batch-feature
937884cf
Merge branch 'image-batch-feature' into image-processor-glpn
a1b681a6
Alias GLPN feature extractor to image processor
b1db434a
Alias feature extractor mixin
f0c14ee1
Resolve merge conflicts
952c2a0a
Resolve merge conflicts
2f0fa0bc
Resolve merge conflicts
e6233cce
Merge branch 'image-processor-glpn' into rename-fe-to-ip-glpn
ddc8cf95
Merge in main
5407de68
Merge branch 'image-transforms-library' into image-processor-mixin
f1cf2286
Merge branch 'image-processor-mixin' into base-image-processor-class
bd0afd69
Merge branch 'base-image-processor-class' into image-batch-feature
a6f69bc9
Merge and resolve conflicts
a7af81fb
Merge branch 'image-processor-glpn' into rename-fe-to-ip-glpn
ad58bd94
Add return_numpy=False flag for resize
affb9450
Merge branch 'image-transforms-library' into image-processor-mixin
5891dd85
Merge branch 'image-processor-mixin' into base-image-processor-class
b66d0f60
Merge branch 'base-image-processor-class' into image-batch-feature
8b73f898
Merge branch 'image-batch-feature' into image-processor-glpn
ae6030c3
Merge branch 'image-processor-glpn' into rename-fe-to-ip-glpn
78bdfb39
Fix up
7a4d22a8
Fix up
994e0402
Use different frameworks safely
42c23bda
Safely import PIL
05c65f68
Call function checking if PIL available
feb9556e
Only import if vision available
a30b007d
Address Sylvain PR comments
fd7b6c7b
Apply suggestions from code review
790c2c6d
Update src/transformers/image_transforms.py
2e929cfd
Update src/transformers/models/glpn/feature_extraction_glpn.py
ff04de34
Merge pull request #25 from amyeroberts/image-processor-mixin
cb1dcd82
Add in docstrings
ae358735
Merge pull request #23 from amyeroberts/image-processor-glpn
4fff2672
Merge pull request #26 from amyeroberts/image-batch-feature
62c6e55d
Merge pull request #24 from amyeroberts/rename-fe-to-ip-glpn
271a09df
Fix TFSwinSelfAttention to have relative position index as non-traina…
b7edea03
Refactor `TFSwinLayer` to increase serving compatibility (#18352)
8cacf308
Add TF prefix to TF-Res test class (#18481)
e385c5a9
Remove py.typed (#18485)
ed4b0595
Fix pipeline tests (#18487)
553be890
Use new huggingface_hub tools for download models (#18438)
cfa16ebe
Fix `test_dbmdz_english` by updating expected values (#18482)
7472b39f
Move cache folder to huggingface/hub for consistency with hf_hub (#18…
35a534af
Update some expected values in `quicktour.mdx` for `resampy 0.3.0` (#…
2c96675d
Forgot one new_ for cache migration
fc879699
disable Onnx test for google/long-t5-tglobal-base (#18454)
e8f57723
Typo reported by Joel Grus on TWTR (#18493)
707c0ff5
Just re-reading the whole doc every couple of months 😬 (#18489)
0ff35a8e
`transformers-cli login` => `huggingface-cli login` (#18490)
1d346569
Add seed setting to image classification example (#18519)
a6101553
[DX fix] Fixing QA pipeline streaming a dataset. (#18516)
2f493d54
Clean up hub (#18497)
80c33f8d
update fsdp docs (#18521)
c6e979f1
Fix compatibility with 1.12 (#17925)
2884397e
Remove debug statement
e9ba674d
Specify en in doc-builder README example (#18526)
dcb1685e
New cache fixes: add safeguard before looking in folders (#18522)
7072f661
unpin resampy (#18527)
c5e228ee
✨ update to use interlibrary links instead of Markdown (#18500)
6952e9b5
Add example of multimodal usage to pipeline tutorial (#18498)
8a18ad9c
[VideoMAE] Add model to doc tests (#18523)
e9c67f74
Update perf_train_gpu_one.mdx (#18532)
7aa5bfd4
Update no_trainer.py scripts to include accelerate gradient accumulat…
5606dba1
Add Spanish translation of converting_tensorflow_models.mdx (#18512)
5b29a580
Spanish translation of summarization.mdx (#15947) (#18477)
defa14cd
Let's not cast them all (#18471)
87271d19
fix: data2vec-vision Onnx ready-made configuration. (#18427)
a9b29687
Add mt5 onnx config (#18394)
24f688fc
Minor update of `run_call_with_unpacked_inputs` (#18541)
c4654370
BART - Fix attention mask device issue on copied models (#18540)
cafb76e4
Adding a new `align_to_words` param to qa pipeline. (#18010)
a25b1b3c
📝 update metric with evaluate (#18535)
fdd9c95c
Restore _init_weights value in no_init_weights (#18504)
5e8a3d44
Clean up comment
ba982711
📝 update documentation build section (#18548)
3a70590f
`bitsandbytes` - `Linear8bitLt` integration into `transformers` model…
09f36ba1
TF: XLA-trainable DeBERTa v2 (#18546)
ca3833e2
Preserve hub-related kwargs in AutoModel.from_pretrained (#18545)
b84379c4
TF Examples Rewrite (#18451)
8d7065ec
Use commit hash to look in cache instead of calling head (#18534)
c9c54204
`pipeline` support for `device="mps"` (or any other string) (#18494)
5d390884
Update philosophy to include other preprocessing classes (#18550)
0544879b
Properly move cache when it is not in default path (#18563)
8b987331
Adds CLIP to models exportable with ONNX (#18515)
c2fc948f
raise atol for MT5OnnxConfig (#18560)
fe29e4c0
fix string (#18568)
793d978d
Segformer TF: fix output size in documentation (#18572)
8aea3311
Fix resizing bug in OWL-ViT (#18573)
db07c44e
Fix LayoutLMv3 documentation (#17932)
5a29d4f3
Skip broken tests
ad4215f9
Change BartLearnedPositionalEmbedding's forward method signature to s…
a272ed0d
german docs translation (#18544)
6d8ab277
Deberta V2: Fix critical trace warnings to allow ONNX export (#18272)
9d87c2de
[FX] _generate_dummy_input supports audio-classification models for l…
1c38f1ac
Fix docstrings with last version of hf-doc-builder styler (#18581)
529ac2bb
![dependabot[bot]](https://avatars.githubusercontent.com/in/29110?v=4&s=60)
Bump nbconvert from 6.0.1 to 6.3.0 in /examples/research_projects/lxm…
8a8a9a1c
Bump nbconvert in /examples/research_projects/visual_bert (#18566)
5a467992
fix owlvit tests, update docstring examples (#18586)
5d1df723
Return the permuted hidden states if return_dict=True (#18578)
f03866fd
Load sharded pt to flax (#18419)
261f4807
Add type hints for ViLT models (#18577)
ff90f496
update doc for perf_train_cpu_many, add intel mpi introduction (#18576)
1e7062af
typos (#18594)
c472b595
FSDP bug fix for `load_state_dict` (#18596)
8cd549f0
Add `TFAutoModelForSemanticSegmentation` to the main `__init__.py` (#…
b0dea996
Generate: validate `model_kwargs` (and catch typos in generate argume…
b93957a9
Supporting seq2seq models for `bitsandbytes` integration (#18579)
b881653f
Add Donut (#18488)
a9a0e18e
Fix URLs (#18604)
089ad230
Update BLOOM parameter counts (#18531)
f1590b23
[doc] fix anchors (#18591)
b2fe78b0
[fsmt] deal with -100 indices in decoder ids (#18592)
c9d8c70d
small change (#18584)
af92441d
Flax Remat for LongT5 (#17994)
2287492b
mac m1 `mps` integration (#18598)
c97b0856
Change scheduled CIs to use torch 1.12.1 (#18644)
ea2c9924
Add checks for some workflow jobs (#18583)
771d6c03
TF: Fix generation repetition penalty with XLA (#18648)
b53ef289
Update longt5.mdx (#18634)
1769f662
Update run_translation_no_trainer.py (#18637)
b2dc2f37
[bnb] Minor modifications (#18631)
ab9d3b44
Examples: add Bloom support for token classification (#18632)
a316ea33
Fix Yolos ONNX export test (#18606)
c6751eac
Fixup
b7046bc8
Fix up
f8a6b878
Resolve conflicts
b6fd4e37
Move PIL default arguments inside function for safe imports
a37bce32
Add image utils to toctree
6ec9dbb2
Update `rescale` method to reflect changes in #18677
76936006
Update docs/source/en/internal/image_processing_utils.mdx
464a4f29
Address Niels PR comments
713e9588
Add normalize method to transforms library
4e60a763
Apply suggestions from code review - remove defaults to None
6ec76ffb
Merge branch 'main' into image-transforms-library
adc0f9d7
Merge branch 'main' into image-transforms-library
df81b6a5
Merge branch 'image-transforms-library' into image-transforms-add-nor…
38a2427e
Fix docstrings and revert to PIL.Image.XXX resampling
48a07a1c
Some more docstrings and PIL.Image tidy up
87852290
Merge branch 'image-transforms-library' into image-transforms-add-nor…
1b2ce563
Reorganise arguments so flags by modifiers
d44fe636
Merge branch 'image-transforms-library' into image-transforms-add-nor…
16991f04
Few last docstring fixes
83330efb
Merge branch 'image-transforms-library' into image-transforms-add-nor…
eba05e21
Merge branch 'main' into image-transforms-library
bad71740
Merge branch 'image-transforms-library' into image-transforms-add-nor…
3191025b
Resolve conflicts with main
3dd6e02c
Add normalize to docs
292f7868
Accept PIL.Image inputs with deprecation warning
7ea393cc
alaradirik
approved these changes
on 2022-10-17
NielsRogge
approved these changes
on 2022-10-17
Update src/transformers/image_transforms.py
0f996616
Update warning to include version
799829bc
Trigger CI - hash clash on doc build
d396382f
Reviewers
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Login to write a write a comment.
Login via GitHub