增加keras-cv模型及训练代码

Keras/keras-cv/.github/.githooks/pre-commit

+#!/bin/sh
+./shell/lint.sh $(git diff HEAD --diff-filter=d --cached --name-only '*.py')

Keras/keras-cv/.github/API_DESIGN.md

+# API Design Guidelines
+In general, KerasCV abides to the
+ [API design guidelines of Keras](https://github.com/keras-team/governance/blob/master/keras_api_design_guidelines.md).
+
+There are a few API guidelines that apply only to KerasCV.  These are discussed
+in this document.
+
+# Label Names
+When working with `bounding_box` and `segmentation_map` labels the 
+abbreviations `bbox` and `segm` are often used.  In KerasCV, we will *not* be 
+using these abbreviations.  This is done to ensure full consistency in our 
+naming convention.  While the team is fond of the abbreviation `bbox`, we are 
+less fond of `segm`.  In order to ensure full consistency, we have decided to
+use the full names for label types in our code base.
+
+# Preprocessing Layers
+## Strength Parameters
+Many augmentation layers take a parameter representing a strength, often called
+`factor`.  When possible, factor values must conform to a the range: `[0, 1]`, with
+1 representing the strongest transformation and 0 representing a no-op transform.
+The strength of an augmentation should scale linearly with this factor.  If needed,
+a transformation can be performed to map to a large value range internally.  If
+this is done, please provide a thorough explanation of the value range semantics in
+the docstring.
+
+Additionally, factors should support both float and tuples as inputs.  If a float is
+passed, such as `factor=0.5`, the layer should default to the range `[0, factor]`.
+
+## BaseImageAugmentationLayer
+When implementing preprocessing, we encourage users to subclass the 
+`keras_cv.layers.preprocessing.BaseImageAugmentationLayer`.  This layer provides
+ a common `call()` method, auto vectorization, and more.  
+
+When subclassing `BaseImageAugmentationLayer`, several methods can overridden:
+
+- `BaseImageAugmentationLayer.augment_image()` must be overridden
+- `augment_label()` allows updates to be made to labels
+- `augment_bounding_box()` allows updates to bounding boxes to be made
+
+[`RandomShear` serves as a canonical example of how to subclass `BaseImageAugmentationLayer`](https://github.com/keras-team/keras-cv/blob/master/keras_cv/layers/preprocessing/random_shear.py)
+
+## Vectorization
+`BaseImageAugmentationLayer` requires you to implement augmentations in an 
+image-wise basis instead of using a vectorized approach.  This design choice 
+was based made on the  results found in the 
+[vectorization\_strategy\_benchmark.py](../benchmarks/vectorization_strategy_benchmark.py) 
+benchmark.
+
+In short, the benchmark shows that making use of `tf.vectorized_map()` performs
+almost identically to a manually vectorized implementation.  As such, we have 
+decided to rely on `tf.vectorized_map()` for performance.
+
+![Results of vectorization strategy benchmark](images/runtime-plot.png)
+
+## Color Based Preprocessing Layers
+Some preprocessing layers in KerasCV perform color based transformations.  This
+includes `RandomBrightness`, `Equalize`, `Solarization`, and more.  
+Preprocessing layers that perform color based transformations make the 
+following assumptions:
+
+- these layers must accept a `value_range`, which is a tuple of numbers.
+- `value_range` must default to `(0, 255)`
+- input images may be of any `dtype`
+
+The decision to support inputs of any `dtype` is made based on the nuance that
+some Keras layers cast user inputs without the user knowing.  For example, if
+`Solarization` expected user inputs to be of type `int`, and a custom layer
+was accidentally casting inputs to `float32`, it would be a bad user experience
+to raise an error asserting that all inputs must be of type `int`.  
+
+New preprocessing layers should be consistent with these decisions.
+
+# Codesamples
+
+- Import symbols from top level namespaces in code samples (usage docstring for example).
+
+Prefer:
+
+```python
+import keras_cv.layers.StochasticDepth
+```
+
+to:
+
+```python
+keras_cv.layers.regularization.stochastic_depth.StochasticDepth
+```
+

Keras/keras-cv/.github/CALL_FOR_CONTRIBUTIONS.md

+# Call For Contributions
+Contributors looking for a task can look at the following list to find an item
+to work on.  Should you decide to contribute a component, please comment on the 
+corresponding GitHub issue that you will be working on the component.  A team 
+member will then follow up by assigning the issue to you.
+
+[There is a contributions welcome label available here](https://github.com/keras-team/keras-cv/issues?page=2&q=is%3Aissue+is%3Aopen+label%3Acontribution-welcome)

Keras/keras-cv/.github/CONTRIBUTING.md

+## How to contribute code
+
+Follow these steps to submit your code contribution.
+
+[You can find a list of issues that we are looking for contributors on here!](https://github.com/keras-team/keras-cv/labels/contribution-welcome)
+
+### Step 1. Open an issue
+
+Before making any changes, we recommend opening an issue (if one doesn't already
+exist) and discussing your proposed changes. This way, we can give you feedback
+and validate the proposed changes.
+
+If your code change involves the fixing of a bug, please include a
+[Colab](https://colab.research.google.com/) notebook that shows
+how to reproduce the broken behavior.
+
+If the changes are minor (simple bug fix or documentation fix), then feel free
+to open a PR without discussion.
+
+### Step 2. Make code changes
+
+To make code changes, you need to fork the repository. You will need to setup a
+development environment and run the unit tests. This is covered in section
+"Setup environment".
+
+If your code change involves introducing a new API change, please see our
+[API Design Guidelines](API_DESIGN.md).
+
+**Notes**
+
+- Make sure to add a new entry to [serialization tests](https://github.com/keras-team/keras-cv/blob/master/keras_cv/layers/serialization_test.py#L37) for new layers.
+
+### Step 3. Create a pull request
+
+Once the change is ready, open a pull request from your branch in your fork to
+the master branch in [keras-team/keras-cv](https://github.com/keras-team/keras-cv).
+
+### Step 4. Sign the Contributor License Agreement
+
+After creating the pull request, you will need to sign the Google CLA agreement.
+The agreement can be found at [https://cla.developers.google.com/clas](https://cla.developers.google.com/clas).
+
+
+### Step 5. Code review
+
+CI tests will automatically be run directly on your pull request.  Their
+status will be reported back via GitHub actions.
+
+There may be
+several rounds of comments and code changes before the pull request gets
+approved by the reviewer.
+
+![Approval from reviewer](https://i.imgur.com/zgRziTt.png)
+
+### Step 6. Merging
+
+Once the pull request is approved, a team member will take care of merging.
+
+## Contributing models
+When contributing new models, please validate model performance by providing training results. You can do this using our existing [ImageNet training script](https://github.com/keras-team/keras-cv/blob/master/examples/training/classification/imagenet/basic_training.py) or by contributing a custom training script of your own (see "Contributing training scripts" below). Training results can be added to the training history log with [this script](https://github.com/keras-team/keras-cv/blob/master/shell/weights/update_training_history.py), or shared with the team via Google Drive (we'll need TensorBoard logs as well as weights). Either way, the KerasCV team will need to upload the weights to our GCS bucket for distribution.
+
+For an initial submission, trained weights do not need to exactly match paper-claimed results. As a baseline, let's shoot for 90% of the paper-claimed ImageNet top-1 accuracy. However, we should strive to improve these weights quickly to at least match paper-claimed results.
+
+## Contributing training scripts
+
+KerasCV is working to include a catalog of high-performing model training scripts for the models included in KerasCV.models and is welcoming contributions for these scripts. These training scripts serve as documentation of good training techniques and will be used to train weights that will be offered in KerasCV models through the package.
+
+The KerasCV team will run submitted training scripts to produce weights for KerasCV, and will attribute strong weights to contributors via a training script ranking system. Stay tuned for more details about that.
+
+Incremental improvements to existing training scripts are welcome, provided that they come with evidence of improved validation performance.
+
+You can also open an issue to add weights for a specific model using a pre-existing script! In your issue, provide your training logs and resulting weights. Specify the arguments that were used to run the script, and provide support for those choices. If your weights beat our current weights, they'll become our default pre-trained weights for your model/task in KerasCV.models!
+
+To contribute a new script, start by opening an issue and tagging @ianjjohnson and @LukeWood to discuss the task, dataset, and/or model for which you'd like to add a script. Once they've taken a look, you can prepare a PR to introduce the new training script.
+
+See [this example script](https://github.com/keras-team/keras-cv/blob/master/examples/training/classification/imagenet/basic_training.py) for training ImageNet classification. Please follow the structure of this training script in contributing your own script. New scripts should either:
+- Train a task for which we don't have a training script already
+- Include a meaningfully different training approach for a given task
+- Introduce a custom training method for a specific model or dataset, based on empirical evidence of efficacy.
+
+When contributing training scripts or proposing runs, please include documentation to support decisions about training including hyperparameter choices. Examples of good documentation would be recent literature or a reference to a hyperparameter search.
+
+Our default training scripts train using ImageNet. Because we cannot distribute this dataset, you will need to modify your dataloading step to load the dataset on your system if you wish to run training yourself. You are also welcome to locally train against a different dataset, provided that you include documentation in your PR supporting the claim that your script will still perform well against ImageNet.
+
+We look forward to delivering great pre-trained models in KerasCV with the help of your contributions!
+
+## Contributing custom ops
+
+We do not plan to accept contributed custom ops due to the maintenance burden that they introduce. If there is a clear need for a specific custom op that should live in KerasCV, please consult the KerasCV team before implementing it, as we expect to reject contributions of custom ops by default.
+
+We currently support only a small handful of ops that run on CPU and are not used at inference time.
+
+If you are updating existing custom ops, you can re-compile the binaries from source using the instructions in the `Tests that require custom ops` section below.
+
+## Setup environment
+
+Setting up your KerasCV development environment requires you to fork the KerasCV repository,
+clone the repository, install dependencies, and execute `python setup.py develop`.
+
+You can achieve this by running the following commands:
+
+```shell
+gh repo fork keras-team/keras-cv --clone --remote
+cd keras-cv
+pip install ".[tests]"
+python setup.py develop
+```
+
+The first line relies on having an installation of [the GitHub CLI](https://github.com/cli/cli).
+
+Following these commands you should be able to run the tests using `pytest keras_cv`.
+Please report any issues running tests following these steps.
+
+Note that this will _not_ install custom ops. If you'd like to install custom ops from source, you can compile the binaries and add them to your local environment manually (requires Bazel):
+
+```shell
+python build_deps/configure.py
+
+bazel build keras_cv/custom_ops:all
+mv bazel-bin/keras_cv/custom_ops/*.so keras_cv/custom_ops
+```
+
+## Run tests
+
+KerasCV is tested using [PyTest](https://docs.pytest.org/en/6.2.x/).
+
+### Run a test file
+
+To run a test file, run `pytest path/to/file` from the root directory of keras\_cv.
+
+### Run a single test case
+
+To run a single test, you can use `-k=<your_regex>`
+to use regular expression to match the test you want to run. For example, you
+can use the following command to run all the tests in `cut_mix_test.py`,
+whose names contain `label`,
+
+```
+pytest keras_cv/layers/preprocessing/cut_mix_test.py -k="label"
+```
+
+### Run all tests
+
+You can run the unit tests for KerasCV by running:
+```
+pytest keras_cv/
+```
+
+### Tests that require custom ops
+For tests that require custom ops, you'll have to compile the custom ops and make them available to your local Python code:
+```shell
+python build_deps/configure.py
+bazel build keras_cv/custom_ops:all
+cp bazel-bin/keras_cv/custom_ops/*.so keras_cv/custom_ops/
+```
+
+Tests which use custom ops are disabled by default, but can be run by setting the environment variable `TEST_CUSTOM_OPS=true`.
+
+## Formatting the Code
+We use `flake8`, `isort` and `black` for code formatting.  You can run
+the following commands manually every time you want to format your code:
+
+- Run `shell/format.sh` to format your code
+- Run `shell/lint.sh` to check the result.
+
+If after running these the CI flow is still failing, try updating `flake8`, `isort` and `black`.
+This can be done by running `pip install --upgrade black`, `pip install --upgrade flake8`, and
+`pip install --upgrade isort`.
+
+Note: The linting checks could be automated activating  
+      pre-commit hooks with `git config core.hooksPath .github/.githooks`
+
+## Community Guidelines
+
+This project follows [Google's Open Source Community Guidelines](https://opensource.google/conduct/).

Keras/keras-cv/.github/ISSUE_TEMPLATE/feature_request.md

+---
+name: Feature request
+about: Suggest an idea for this project
+title: ''
+labels: ''
+assignees: ''
+
+---
+
+**Short Description**
+<!---
+A clear and concise description of what the feature is.
+-->
+
+**Papers**
+<!---
+Include citation counts if possible.
+-->
+
+**Existing Implementations**
+<!---
+Link to existing implementations.  TensorFlow implementations are preferred.
+-->
+
+**Other Information**
+

Keras/keras-cv/.github/ISSUE_TEMPLATE/other.md

+---
+name: Other issue template
+about: Describe this issue template's purpose here.
+title: ''
+labels: ''
+assignees: ''
+
+---
+
+

Keras/keras-cv/.github/PULL_REQUEST_TEMPLATE.md

+# What does this PR do?
+
+<!--
+Congratulations! You've made it this far! You're not quite done yet though.
+
+Once merged, your PR is going to appear in the release notes with the title you set, so make sure it's a great title that fully reflects the extent of your awesome contribution.
+
+Then, please replace this with a description of the change and which issue is fixed (if applicable). Please also include relevant motivation and context. List any dependencies (if any) that are required for this change.
+
+Once you're done, someone will review your PR shortly (see the section "Who can review?" below to tag some potential reviewers). They may suggest changes to make the code even better. If no one reviewed your PR after a week has passed, don't hesitate to post a new comment @-mentioning the same persons --sometimes notifications get lost.
+-->
+
+<!-- Remove if not applicable -->
+
+Fixes # (issue)
+
+
+## Before submitting
+- [ ] This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
+- [ ] Did you read the [contributor guideline](https://github.com/keras-team/keras-cv/blob/master/.github/CONTRIBUTING.md),
+      Pull Request section?
+- [ ] Was this discussed/approved via a Github issue? Please add a link
+      to it if that's the case.
+- [ ] Did you write any new necessary tests?
+- [ ] If this adds a new model, can you run a few training steps on TPU in Colab to ensure that no XLA incompatible OP are used?
+
+## Who can review?
+
+Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
+members/contributors who may be interested in your PR.
+
+<!-- 
+Feel free to tag @LukeWood and @tanzhenyu in your reviews.
+-->
+
+<!--
+This PR template is copied and modified from here:
+https://github.com/huggingface/transformers/blob/main/.github/PULL_REQUEST_TEMPLATE.md
+-->

Keras/keras-cv/.github/ROADMAP.md

+# Roadmap
+The team will release 2 quarters of roadmap in advance so contributors will know
+what we are working on and be better aligned when creating PRs.  
+As an exception, widely used backbones are always welcome. Contributors can search for `contribution-welcome` label on github.
+The team will release one minor version upgrade each quarter, or whenever a new task is officially supported.
+
+2023 Q1:
+- 3D Object Detection task, the team focuses on providing a point cloud based model (specific model TBD)
+- OCR task, the team opens up contribution.
+- Keypoint and Action Detection task, the team opens up contribution for Simple Pose, i3d etc.
+- Open Vocabulary Object Detection task, built on top of 2D Object Detection.
+
+
+2022 Q4:
+- 2D Semantic Segmentation task, the team focuses on DeepLabV3 and opens up contribution
+  for UNet/FCN.
+- 3D Object Detection task, the team focuses on providing point cloud data pipeline.
+- 2D Instance Segmentation task, the team focuses on MaskRCNN meta arch built on top of FasterRCNN.
+- Text-to-image task, more specifically built on top of StableDiffusion for image impainting and text-to-video.
+- Refactoring data augmentation layers:
+    - refactoring bounding box input format
+    - `RaggedTensor` support for data augmentation layers, contribution welcome.
+
+
+2022 Q3:
+- 2D Object Detection task, the team focuses on RetinaNet and FasterRCNN meta arch, and opens
+  up contribution for YOLOv3, YOLOx and CenterNet.
+- StableDiffusion model with pretrained weights

Keras/keras-cv/.github/workflows/actions.yml

+name: Tests
+
+on:
+  push:
+  pull_request:
+  release:
+    types: [created]
+jobs:
+  test:
+    name: Test the code
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [windows-latest, ubuntu-latest]
+    steps:
+    - uses: actions/checkout@v2
+    - name: Set up Python 3.7
+      uses: actions/setup-python@v1
+      with:
+        python-version: 3.7
+    - name: Get pip cache dir
+      id: pip-cache
+      run: |
+        python -m pip install --upgrade pip setuptools
+        echo "::set-output name=dir::$(pip cache dir)"
+    - name: pip cache
+      uses: actions/cache@v2
+      with:
+        path: ${{ steps.pip-cache.outputs.dir }}
+        key: ${{ runner.os }}-pip-${{ hashFiles('setup.py') }}
+        restore-keys: |
+          ${{ runner.os }}-pip-
+    - name: Install dependencies
+      run: |
+        pip install tensorflow-cpu==2.10.0
+        pip install -e ".[tests]" --progress-bar off --upgrade
+    - name: Build custom ops for tests
+      run: |
+        python build_deps/configure.py
+        bazel build keras_cv/custom_ops:all
+        cp bazel-bin/keras_cv/custom_ops/*.so keras_cv/custom_ops/
+    - name: Test with pytest
+      env:
+        TEST_CUSTOM_OPS: true
+      run: |
+        pytest keras_cv/ --ignore keras_cv/models
+  format:
+    name: Check the code format
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v2
+    - name: Set up Python 3.7
+      uses: actions/setup-python@v1
+      with:
+        python-version: 3.7
+    - name: Get pip cache dir
+      id: pip-cache
+      run: |
+        python -m pip install --upgrade pip setuptools
+        echo "::set-output name=dir::$(pip cache dir)"
+    - name: pip cache
+      uses: actions/cache@v2
+      with:
+        path: ${{ steps.pip-cache.outputs.dir }}
+        key: ${{ runner.os }}-pip-${{ hashFiles('setup.py') }}
+        restore-keys: |
+          ${{ runner.os }}-pip-
+    - name: Install dependencies
+      run: |
+        pip install tensorflow==2.10.0
+        pip install -e ".[tests]" --progress-bar off --upgrade
+    - name: Lint
+      run: bash shell/lint.sh
+  deploy:
+    needs: [test, format]
+    if: github.event_name == 'release' && github.event.action == 'created'
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v2
+    - name: Set up Python
+      uses: actions/setup-python@v1
+      with:
+        python-version: 3.7
+    - name: Install dependencies
+      run: |
+        pip install tensorflow==2.10.0
+        python -m pip install --upgrade pip setuptools wheel twine
+    - name: Build and publish
+      env:
+        TWINE_USERNAME: ${{ secrets.PYPI_USERNAME }}
+        TWINE_PASSWORD: ${{ secrets.PYPI_PASSWORD }}
+      run: |
+        python setup.py sdist bdist_wheel
+        twine upload dist/*

Keras/keras-cv/.github/workflows/devcontainer.yml

+name: 'devcontainer' 
+on: # rebuild any PRs and main branch changes
+  pull_request:
+    paths:
+      - 'setup.*'
+      - '.devcontainer/**'
+  push:
+    paths:
+      - 'setup.*'
+      - '.devcontainer/**'
+
+
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+
+      - name: Checkout (GitHub)
+        uses: actions/checkout@v2
+
+      - name: Build and run dev container task
+        uses: devcontainers/ci@v0.2
+        with:
+          runCmd: pytest keras_cv/ --ignore keras_cv/models

Keras/keras-cv/.gitignore

+keras_cv.egg-info/
+wandb/
+logs/
+dist/
+.DS_Store
+build/
+*.swp
+.idea
+*.pyc
+.pytest_cache
+*.egg-info
+__pycache__/
+
+#VS Code files and container
+.vscode/
+.devcontainer/
+.coverage

Keras/keras-cv/BUILD.bazel

+sh_binary(
+    name = "build_pip_pkg",
+    srcs = ["build_deps/build_pip_pkg.sh"],
+    data = [
+        "LICENSE",
+        "MANIFEST.in",
+        "README.md",
+        "setup.py",
+        "//keras_cv",
+    ],
+)

Keras/keras-cv/LICENSE

+Copyright 2022 The KerasCV Authors. All rights reserved.
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2015, The TensorFlow Authors.
+
+   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.

Keras/keras-cv/MANIFEST.in

+recursive-include keras_cv/custom_ops *.so

Keras/keras-cv/README.md

+
+Introduction：
+            
+Keras-CV源码及训练代码示例，TensorFlow版本为2.9，python版本支持3.7、3.8、3.9.
+支持的网络有 ResNet101、ResNet50V2、RegNetX064、DarkNet53、DenseNet121、EfficientNetV2S
+网络模型代码位于keras-cv/models.训练代码basic_training.py位于examples/training/classification/imagenet/
+使用tfrecord格式的Imagenet数据集.相关网络的训练情况及参数设置位于examples/training/classification/imagenet/
+rain_history.json
+
+Install
+
+安装DTK版本的TensorFlow2.9 pip install tensorflow-2.9.0+git28e158bd.dtk22042-cp37-cp37m-linux_x86_64.whl
+执行python setup.py build
+   python setup.py install
+       
+Train   
+训练脚本位于examples/training/classification/imagenet/trian_keras_cv.sh 注意basic_training.py中的分布式设置为
+默认使用所有检测到的加速卡设备，若要指定几张加速卡参与训练使用export HIP_VISIBLE_DEVICES=x1，x2，x3来指定.
\ No newline at end of file

Keras/keras-cv/benchmarks/metrics/coco/mean_average_precision_bucket_performance.py

+import math
+import random
+import time
+
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+import seaborn as sns
+import tensorflow as tf
+
+import keras_cv
+from keras_cv.metrics import coco
+
+
+def produce_random_data(include_confidence=False, num_images=128, classes=20):
+    """Generates a fake list of bounding boxes for use in this test.
+
+    Returns:
+      a tensor list of size [128, 25, 5/6].  This represents 128 images, 25 bboxes
+        and 5/6 dimensions to represent each bbox depending on if confidence is
+        set.
+    """
+    images = []
+    for _ in range(num_images):
+        num_boxes = math.floor(25 * random.uniform(0, 1))
+        classes_in_image = np.floor(np.random.rand(num_boxes, 1) * classes)
+        bboxes = np.random.rand(num_boxes, 4)
+        boxes = np.concatenate([bboxes, classes_in_image], axis=-1)
+        if include_confidence:
+            confidence = np.random.rand(num_boxes, 1)
+            boxes = np.concatenate([boxes, confidence], axis=-1)
+        images.append(
+            keras_cv.utils.bounding_box.xywh_to_corners(
+                tf.constant(boxes, dtype=tf.float32)
+            )
+        )
+
+    images = [
+        keras_cv.bounding_box.pad_batch_to_shape(x, [25, images[0].shape[1]])
+        for x in images
+    ]
+    return tf.stack(images, axis=0)
+
+
+y_true = produce_random_data()
+y_pred = produce_random_data(include_confidence=True)
+class_ids = list(range(20))
+
+bucket_values = [500, 1000, 2000, 3500, 5000, 7500, 10000]
+
+update_state_runtimes = []
+result_runtimes = []
+end_to_end_runtimes = []
+
+for buckets in bucket_values:
+    metric = coco.COCOMeanAveragePrecision(class_ids, num_buckets=buckets)
+    # warm up
+    metric.update_state(y_true, y_pred)
+    metric.result()
+
+    start = time.time()
+    metric.update_state(y_true, y_pred)
+    update_state_done = time.time()
+    r = metric.result()
+    end = time.time()
+
+    update_state_runtimes.append(update_state_done - start)
+    result_runtimes.append(end - update_state_done)
+    end_to_end_runtimes.append(end - start)
+
+    print("end_to_end_runtimes", end_to_end_runtimes)
+
+data = pd.DataFrame(
+    {
+        "bucket_values": bucket_values,
+        "update_state_runtimes": update_state_runtimes,
+        "result_runtimes": result_runtimes,
+        "end_to_end_runtimes": end_to_end_runtimes,
+    }
+)
+
+sns.lineplot(data=data, x="bucket_values", y="update_state_runtimes")
+plt.xlabel("Number of Confidence Buckets")
+plt.ylabel("update_state() runtime (seconds)")
+plt.title("Runtime of update_state()")
+plt.show()
+
+sns.lineplot(data=data, x="bucket_values", y="result_runtimes")
+plt.xlabel("Number of Confidence Buckets")
+plt.ylabel("result() runtime (seconds)")
+plt.title("Runtime of result()")
+plt.show()
+
+sns.lineplot(data=data, x="bucket_values", y="end_to_end_runtimes")
+plt.xlabel("Number of Confidence Buckets")
+plt.ylabel("End to end runtime (seconds)")
+plt.title("Runtimes of update_state() followed by result()")
+plt.show()

Keras/keras-cv/benchmarks/metrics/coco/mean_average_precision_performance.py

+import math
+import random
+import time
+
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+import seaborn as sns
+import tensorflow as tf
+
+import keras_cv
+from keras_cv.metrics import coco
+
+
+def produce_random_data(include_confidence=False, num_images=128, classes=20):
+    """Generates a fake list of bounding boxes for use in this test.
+
+    Returns:
+      a tensor list of size [128, 25, 5/6].  This represents 128 images, 25 bboxes
+        and 5/6 dimensions to represent each bbox depending on if confidence is
+        set.
+    """
+    images = []
+    for _ in range(num_images):
+        num_boxes = math.floor(25 * random.uniform(0, 1))
+        classes_in_image = np.floor(np.random.rand(num_boxes, 1) * classes)
+        bboxes = np.random.rand(num_boxes, 4)
+        boxes = np.concatenate([bboxes, classes_in_image], axis=-1)
+        if include_confidence:
+            confidence = np.random.rand(num_boxes, 1)
+            boxes = np.concatenate([boxes, confidence], axis=-1)
+        images.append(
+            keras_cv.utils.bounding_box.xywh_to_corners(
+                tf.constant(boxes, dtype=tf.float32)
+            )
+        )
+
+    images = [
+        keras_cv.bounding_box.pad_batch_to_shape(x, [25, images[0].shape[1]])
+        for x in images
+    ]
+    return tf.stack(images, axis=0)
+
+
+y_true = produce_random_data()
+y_pred = produce_random_data(include_confidence=True)
+class_ids = list(range(20))
+
+n_images = [128, 256, 512, 512 + 256, 1024]
+
+update_state_runtimes = []
+result_runtimes = []
+end_to_end_runtimes = []
+
+for images in n_images:
+    y_true = produce_random_data(num_images=images)
+    y_pred = produce_random_data(num_images=images, include_confidence=True)
+    metric = coco.COCOMeanAveragePrecision(class_ids)
+    # warm up
+    metric.update_state(y_true, y_pred)
+    metric.result()
+
+    start = time.time()
+    metric.update_state(y_true, y_pred)
+    update_state_done = time.time()
+    r = metric.result()
+    end = time.time()
+
+    update_state_runtimes.append(update_state_done - start)
+    result_runtimes.append(end - update_state_done)
+    end_to_end_runtimes.append(end - start)
+
+    print("end_to_end_runtimes", end_to_end_runtimes)
+
+data = pd.DataFrame(
+    {
+        "n_images": n_images,
+        "update_state_runtimes": update_state_runtimes,
+        "result_runtimes": result_runtimes,
+        "end_to_end_runtimes": end_to_end_runtimes,
+    }
+)
+
+sns.lineplot(data=data, x="n_images", y="update_state_runtimes")
+plt.xlabel("Number of Images")
+plt.ylabel("update_state() runtime (seconds)")
+plt.title("Runtime of update_state()")
+plt.show()
+
+sns.lineplot(data=data, x="n_images", y="result_runtimes")
+plt.xlabel("Number of Images")
+plt.ylabel("result() runtime (seconds)")
+plt.title("Runtime of result()")
+plt.show()
+
+sns.lineplot(data=data, x="n_images", y="end_to_end_runtimes")
+plt.xlabel("Number of Images")
+plt.ylabel("End to end runtime (seconds)")
+plt.title("Runtimes of update_state() followed by result()")
+plt.show()

Keras/keras-cv/benchmarks/metrics/coco/recall_performance.py

+import math
+import random
+import time
+
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+import seaborn as sns
+import tensorflow as tf
+
+import keras_cv
+from keras_cv.metrics import coco
+
+
+def produce_random_data(include_confidence=False, num_images=128, classes=20):
+    """Generates a fake list of bounding boxes for use in this test.
+
+    Returns:
+      a tensor list of size [128, 25, 5/6].  This represents 128 images, 25 bboxes
+        and 5/6 dimensions to represent each bbox depending on if confidence is
+        set.
+    """
+    images = []
+    for _ in range(num_images):
+        num_boxes = math.floor(25 * random.uniform(0, 1))
+        classes_in_image = np.floor(np.random.rand(num_boxes, 1) * classes)
+        bboxes = np.random.rand(num_boxes, 4)
+        boxes = np.concatenate([bboxes, classes_in_image], axis=-1)
+        if include_confidence:
+            confidence = np.random.rand(num_boxes, 1)
+            boxes = np.concatenate([boxes, confidence], axis=-1)
+        images.append(
+            keras_cv.utils.bounding_box.xywh_to_corners(
+                tf.constant(boxes, dtype=tf.float32)
+            )
+        )
+
+    images = [
+        keras_cv.bounding_box.pad_batch_to_shape(x, [25, images[0].shape[1]])
+        for x in images
+    ]
+    return tf.stack(images, axis=0)
+
+
+y_true = produce_random_data()
+y_pred = produce_random_data(include_confidence=True)
+class_ids = list(range(20))
+
+n_images = [128, 256, 512, 512 + 256, 1024]
+
+
+update_state_runtimes = []
+result_runtimes = []
+end_to_end_runtimes = []
+
+for images in n_images:
+    y_true = produce_random_data(num_images=images)
+    y_pred = produce_random_data(num_images=images, include_confidence=True)
+    metric = coco.COCORecall(class_ids)
+    # warm up
+    metric.update_state(y_true, y_pred)
+    metric.result()
+
+    start = time.time()
+    metric.update_state(y_true, y_pred)
+    update_state_done = time.time()
+    r = metric.result()
+    end = time.time()
+
+    update_state_runtimes.append(update_state_done - start)
+    result_runtimes.append(end - update_state_done)
+    end_to_end_runtimes.append(end - start)
+
+    print("end_to_end_runtimes", end_to_end_runtimes)
+
+
+data = pd.DataFrame(
+    {
+        "n_images": n_images,
+        "update_state_runtimes": update_state_runtimes,
+        "result_runtimes": result_runtimes,
+        "end_to_end_runtimes": end_to_end_runtimes,
+    }
+)
+
+sns.lineplot(data=data, x="n_images", y="update_state_runtimes")
+plt.xlabel("Number of Images")
+plt.ylabel("update_state() runtime (seconds)")
+plt.title("Runtime of update_state()")
+plt.show()
+
+sns.lineplot(data=data, x="n_images", y="result_runtimes")
+plt.xlabel("Number of Images")
+plt.ylabel("result() runtime (seconds)")
+plt.title("Runtime of result()")
+plt.show()
+
+sns.lineplot(data=data, x="n_images", y="end_to_end_runtimes")
+plt.xlabel("Number of Images")
+plt.ylabel("End to end runtime (seconds)")
+plt.title("Runtimes of update_state() followed by result()")
+plt.show()