# An example of Compose? **URL:** \ **Category:** General Discussion\ **Tags:** addons\ **Created:** [June 6, 2021, 3:55pm UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313 "2021-06-06T15:55:45Z")\ **Posts on this page:** 11\ **Page:** 1
**Author:** ![Sayak\_Paul](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/sayak_paul/32/13002_2.png) [@Sayak\_Paul](https://discuss.ai.google.dev/u/Sayak_Paul)\ **Post date:** [June 6, 2021, 3:55pm UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/1 "2021-06-06T15:55:45Z")
@Bhack could you provide a minimal example of `tfa.image.compose_transforms()`? > **[tfa.image.compose\_transforms  |  TensorFlow Addons](https://www.tensorflow.org/addons/api_docs/python/tfa/image/compose_transforms)** > > Composes the transforms tensors. ---
**Author:** ![Bhack](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/bhack/32/6388_2.png) [@Bhack](https://discuss.ai.google.dev/u/Bhack)\ **Post date:** [June 6, 2021, 8:06pm UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/2 "2021-06-06T20:06:23Z")
Here you have rotation + translation > ---
**Author:** ![Sayak\_Paul](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/sayak_paul/32/13002_2.png) [@Sayak\_Paul](https://discuss.ai.google.dev/u/Sayak_Paul)\ **Post date:** [June 7, 2021, 1:48am UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/3 "2021-06-07T01:48:39Z")
Thanks but in practice what are we supposed to be providing to `compose_transforms()`? Something like the following? ```auto compose_transforms([tf.image.rot90, tf.image.random_flip_left_right]) ``` If so, how it’s supposed to be applied to an image or a batch of images? From the documentation and also the test-case shared above, these questions are not clear. ---
**Author:** ![Bhack](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/bhack/32/6388_2.png) [@Bhack](https://discuss.ai.google.dev/u/Bhack)\ **Post date:** [June 7, 2021, 10:06am UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/4 "2021-06-07T10:06:00Z")
Internally it is just a matmul loop over single projective tranfroms matrix in the list (or batched): > ---
**Author:** ![Sayak\_Paul](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/sayak_paul/32/13002_2.png) [@Sayak\_Paul](https://discuss.ai.google.dev/u/Sayak_Paul)\ **Post date:** [June 7, 2021, 10:27am UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/5 "2021-06-07T10:27:17Z")
Sadly, I am still not clear as to how to use it with a dataset or even a single image having a leading batch dimension. ---
**Author:** ![Bhack](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/bhack/32/6388_2.png) [@Bhack](https://discuss.ai.google.dev/u/Bhack)\ **Post date:** [June 7, 2021, 10:42am UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/6 "2021-06-07T10:42:43Z")
That method is just going to compose projective transformations. The real transformation is in `transform` that is still quite duplicated in addons cause the `transform` in TF is still a private API (but the transform Matrix format is the same). See the thread starting at [https://tensorflow-prod.ospodiscourse.com/t/cannot-run-model-fit-locally-help/1648/9](https://tensorflow-prod.ospodiscourse.com/t/cannot-run-model-fit-locally-help/1648/9) ---
**Author:** ![Sayak\_Paul](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/sayak_paul/32/13002_2.png) [@Sayak\_Paul](https://discuss.ai.google.dev/u/Sayak_Paul)\ **Post date:** [June 7, 2021, 10:51am UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/7 "2021-06-07T10:51:28Z")
Here’s a shot example applying the composition on a standard image: > **[Google Colab](https://colab.research.google.com/gist/sayakpaul/6e39e972ba0075801caa907f3fa4eeba/scratchpad.ipynb)** ---
**Author:** ![Bhack](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/bhack/32/6388_2.png) [@Bhack](https://discuss.ai.google.dev/u/Bhack)\ **Post date:** [June 7, 2021, 11:53am UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/8 "2021-06-07T11:53:53Z")
This is the same with internals ```python from tensorflow.python.keras.layers.preprocessing.image_preprocessing import transform .... image_transformed = transform(tf.expand_dims(image, axis=0),composed) ``` ---
**Author:** ![Bhack](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/bhack/32/6388_2.png) [@Bhack](https://discuss.ai.google.dev/u/Bhack)\ **Post date:** [June 7, 2021, 12:01pm UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/9 "2021-06-07T12:01:07Z")
E.g. the internal one was used in transformations that support models in model gardens SIG. [https://github.com/tensorflow/models/blob/master/official/vision/beta/ops/augment.py](https://github.com/tensorflow/models/blob/master/official/vision/beta/ops/augment.py) These transformations are available for users in [https://pypi.org/project/tf-models-official](https://pypi.org/project/tf-models-official) Most of the transformations available in Addons I think that are duplicated now. ---
**Author:** ![Sayak\_Paul](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/sayak_paul/32/13002_2.png) [@Sayak\_Paul](https://discuss.ai.google.dev/u/Sayak_Paul)\ **Post date:** [June 7, 2021, 12:07pm UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/10 "2021-06-07T12:07:34Z")
Cool. But I think we both could agree that the documentation does not clearly reflect on these points. I guess we cannot expect the majority of users to dig into the source code and figure their way out into using these. ---
**Author:** ![Bhack](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/bhack/32/6388_2.png) [@Bhack](https://discuss.ai.google.dev/u/Bhack)\ **Post date:** [June 7, 2021, 12:29pm UTC](https://discuss.ai.google.dev/t/an-example-of-compose/31313/11 "2021-06-07T12:29:01Z")
Yes probably some documentation is missing but I think that we can build a clear path and better documentation or tutorials when we will unify a little bit the image processing API that currently it is quite duplicated or sparse in the ecosystem (also generally we cannot promote private API to the end users like `from tensorflow.python`). The modern history of this in Addons/Ecosystem it is already quite old: > > > \*\*Describe the feature and the current behavior/state.\*\* > RandAugment and AutoAu…gment are both policies for enhanced image preprocessing that are included in EfficientNet, but are still using \`tf.contrib\`. > > https://github.com/tensorflow/tpu/blob/master/models/official/efficientnet/autoaugment.py > > The only \`tf.contrib\` image operations that they use, however, are \[rotate\](https://www.tensorflow.org/addons/api\_docs/python/tfa/image/rotate), \[translate\](https://www.tensorflow.org/addons/api\_docs/python/tfa/image/translate) and \[transform\](https://www.tensorflow.org/addons/api\_docs/python/tfa/image/transform) - all of which have been included in TensorFlow Addons. > > \*\*Relevant information\*\* > \- Are you willing to contribute it (yes/no): > No, but am hoping that someone from the community will pick it up (potentially a Google Summer of Code student)? > > \- Are you willing to maintain it going forward? (yes/no): > Yes > > \- Is there a relevant academic paper? (if so, where): > AutoAugment Reference: https://arxiv.org/abs/1805.09501 > RandAugment Reference: https://arxiv.org/abs/1909.13719 > > \- Is there already an implementation in another framework? (if so, where): > See link above; this would be a standard migration from \`tf.contrib\`. > > \- Was it part of tf.contrib? (if so, where): > Yes > > \*\*Which API type would this fall under (layer, metric, optimizer, etc.)\*\* > Image > > \*\*Who will benefit with this feature?\*\* > Anyone doing image preprocessing, especially for EfficientNet. > > > Part 2: https://github.com/tensorflow/addons/issues/1353 > > As part of #1226, re…levant image ops need to be added to \`tfa.image\`. List of ops tracked in this issue (tick the op only once the corresponding PR is \*merged\*): > > \- \[x\] Blend > \- \[x\] CutOut > \- \[\] Identity > \- \[x\] Translate X, Y > \- \[x\] Shear X, Y > \- \[\] Solarize Ops > \- \[\] Posterize > \- \[\] AutoContrast > \- \[x\] Equalize > \- \[\] Invert > \- \[\] Color > \- \[\] Brightness > \- \[x\] Sharpness > > Reference: https://github.com/tensorflow/tpu/blob/master/models/official/efficientnet/autoaugment.py