Upload 74 files

.gitattributes

@@ -36,3 +36,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 GroundingDINO/.asset/GD_GLIGEN.png filter=lfs diff=lfs merge=lfs -text
 GroundingDINO/.asset/GD_SD.png filter=lfs diff=lfs merge=lfs -text
 GroundingDINO/.asset/hero_figure.png filter=lfs diff=lfs merge=lfs -text
+segment_anything/assets/masks1.png filter=lfs diff=lfs merge=lfs -text
+segment_anything/assets/notebook2.png filter=lfs diff=lfs merge=lfs -text
+segment_anything/assets/poster.png filter=lfs diff=lfs merge=lfs -text

segment_anything/.flake8

@@ -0,0 +1,7 @@
+[flake8]
+ignore = W503, E203, E221, C901, C408, E741, C407, B017, F811, C101, EXE001, EXE002
+max-line-length = 100
+max-complexity = 18
+select = B,C,E,F,W,T4,B9
+per-file-ignores =
+  **/__init__.py:F401,F403,E402

segment_anything/CODE_OF_CONDUCT.md

@@ -0,0 +1,80 @@
+# Code of Conduct
+
+## Our Pledge
+
+In the interest of fostering an open and welcoming environment, we as
+contributors and maintainers pledge to make participation in our project and
+our community a harassment-free experience for everyone, regardless of age, body
+size, disability, ethnicity, sex characteristics, gender identity and expression,
+level of experience, education, socio-economic status, nationality, personal
+appearance, race, religion, or sexual identity and orientation.
+
+## Our Standards
+
+Examples of behavior that contributes to creating a positive environment
+include:
+
+* Using welcoming and inclusive language
+* Being respectful of differing viewpoints and experiences
+* Gracefully accepting constructive criticism
+* Focusing on what is best for the community
+* Showing empathy towards other community members
+
+Examples of unacceptable behavior by participants include:
+
+* The use of sexualized language or imagery and unwelcome sexual attention or
+  advances
+* Trolling, insulting/derogatory comments, and personal or political attacks
+* Public or private harassment
+* Publishing others' private information, such as a physical or electronic
+  address, without explicit permission
+* Other conduct which could reasonably be considered inappropriate in a
+  professional setting
+
+## Our Responsibilities
+
+Project maintainers are responsible for clarifying the standards of acceptable
+behavior and are expected to take appropriate and fair corrective action in
+response to any instances of unacceptable behavior.
+
+Project maintainers have the right and responsibility to remove, edit, or
+reject comments, commits, code, wiki edits, issues, and other contributions
+that are not aligned to this Code of Conduct, or to ban temporarily or
+permanently any contributor for other behaviors that they deem inappropriate,
+threatening, offensive, or harmful.
+
+## Scope
+
+This Code of Conduct applies within all project spaces, and it also applies when
+an individual is representing the project or its community in public spaces.
+Examples of representing a project or community include using an official
+project e-mail address, posting via an official social media account, or acting
+as an appointed representative at an online or offline event. Representation of
+a project may be further defined and clarified by project maintainers.
+
+This Code of Conduct also applies outside the project spaces when there is a
+reasonable belief that an individual's behavior may have a negative impact on
+the project or its community.
+
+## Enforcement
+
+Instances of abusive, harassing, or otherwise unacceptable behavior may be
+reported by contacting the project team at <opensource-conduct@fb.com>. All
+complaints will be reviewed and investigated and will result in a response that
+is deemed necessary and appropriate to the circumstances. The project team is
+obligated to maintain confidentiality with regard to the reporter of an incident.
+Further details of specific enforcement policies may be posted separately.
+
+Project maintainers who do not follow or enforce the Code of Conduct in good
+faith may face temporary or permanent repercussions as determined by other
+members of the project's leadership.
+
+## Attribution
+
+This Code of Conduct is adapted from the [Contributor Covenant][homepage], version 1.4,
+available at https://www.contributor-covenant.org/version/1/4/code-of-conduct.html
+
+[homepage]: https://www.contributor-covenant.org
+
+For answers to common questions about this code of conduct, see
+https://www.contributor-covenant.org/faq

segment_anything/CONTRIBUTING.md

@@ -0,0 +1,31 @@
+# Contributing to segment-anything
+We want to make contributing to this project as easy and transparent as
+possible.
+
+## Pull Requests
+We actively welcome your pull requests.
+
+1. Fork the repo and create your branch from `main`.
+2. If you've added code that should be tested, add tests.
+3. If you've changed APIs, update the documentation.
+4. Ensure the test suite passes.
+5. Make sure your code lints, using the `linter.sh` script in the project's root directory. Linting requires `black==23.*`, `isort==5.12.0`, `flake8`, and `mypy`.
+6. If you haven't already, complete the Contributor License Agreement ("CLA").
+
+## Contributor License Agreement ("CLA")
+In order to accept your pull request, we need you to submit a CLA. You only need
+to do this once to work on any of Facebook's open source projects.
+
+Complete your CLA here: <https://code.facebook.com/cla>
+
+## Issues
+We use GitHub issues to track public bugs. Please ensure your description is
+clear and has sufficient instructions to be able to reproduce the issue.
+
+Facebook has a [bounty program](https://www.facebook.com/whitehat/) for the safe
+disclosure of security bugs. In those cases, please go through the process
+outlined on that page and do not file a public issue.
+
+## License
+By contributing to segment-anything, you agree that your contributions will be licensed
+under the LICENSE file in the root directory of this source tree.

segment_anything/LICENSE

@@ -0,0 +1,201 @@
+                                 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 [yyyy] [name of copyright owner]
+
+   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.

segment_anything/README.md

@@ -0,0 +1,107 @@
+# Segment Anything
+
+**[Meta AI Research, FAIR](https://ai.facebook.com/research/)**
+
+[Alexander Kirillov](https://alexander-kirillov.github.io/), [Eric Mintun](https://ericmintun.github.io/), [Nikhila Ravi](https://nikhilaravi.com/), [Hanzi Mao](https://hanzimao.me/), Chloe Rolland, Laura Gustafson, [Tete Xiao](https://tetexiao.com), [Spencer Whitehead](https://www.spencerwhitehead.com/), Alex Berg, Wan-Yen Lo, [Piotr Dollar](https://pdollar.github.io/), [Ross Girshick](https://www.rossgirshick.info/)
+
+[[`Paper`](https://ai.facebook.com/research/publications/segment-anything/)] [[`Project`](https://segment-anything.com/)] [[`Demo`](https://segment-anything.com/demo)] [[`Dataset`](https://segment-anything.com/dataset/index.html)] [[`Blog`](https://ai.facebook.com/blog/segment-anything-foundation-model-image-segmentation/)]
+
+![SAM design](assets/model_diagram.png?raw=true)
+
+The **Segment Anything Model (SAM)** produces high quality object masks from input prompts such as points or boxes, and it can be used to generate masks for all objects in an image. It has been trained on a [dataset](https://segment-anything.com/dataset/index.html) of 11 million images and 1.1 billion masks, and has strong zero-shot performance on a variety of segmentation tasks.
+
+<p float="left">
+  <img src="assets/masks1.png?raw=true" width="37.25%" />
+  <img src="assets/masks2.jpg?raw=true" width="61.5%" /> 
+</p>
+
+## Installation
+
+The code requires `python>=3.8`, as well as `pytorch>=1.7` and `torchvision>=0.8`. Please follow the instructions [here](https://pytorch.org/get-started/locally/) to install both PyTorch and TorchVision dependencies. Installing both PyTorch and TorchVision with CUDA support is strongly recommended.
+
+Install Segment Anything:
+
+```
+pip install git+https://github.com/facebookresearch/segment-anything.git
+```
+
+or clone the repository locally and install with
+
+```
+git clone git@github.com:facebookresearch/segment-anything.git
+cd segment-anything; pip install -e .
+```
+
+The following optional dependencies are necessary for mask post-processing, saving masks in COCO format, the example notebooks, and exporting the model in ONNX format. `jupyter` is also required to run the example notebooks.
+```
+pip install opencv-python pycocotools matplotlib onnxruntime onnx
+```
+
+
+## <a name="GettingStarted"></a>Getting Started
+
+First download a [model checkpoint](#model-checkpoints). Then the model can be used in just a few lines to get masks from a given prompt:
+
+```
+from segment_anything import build_sam, SamPredictor 
+predictor = SamPredictor(build_sam(checkpoint="</path/to/model.pth>"))
+predictor.set_image(<your_image>)
+masks, _, _ = predictor.predict(<input_prompts>)
+```
+
+or generate masks for an entire image:
+
+```
+from segment_anything import build_sam, SamAutomaticMaskGenerator
+mask_generator = SamAutomaticMaskGenerator(build_sam(checkpoint="</path/to/model.pth>"))
+masks = mask_generator_generate(<your_image>)
+```
+
+Additionally, masks can be generated for images from the command line:
+
+```
+python scripts/amg.py --checkpoint <path/to/sam/checkpoint> --input <image_or_folder> --output <output_directory>
+```
+
+See the examples notebooks on [using SAM with prompts](/notebooks/predictor_example.ipynb) and [automatically generating masks](/notebooks/automatic_mask_generator_example.ipynb) for more details.
+
+<p float="left">
+  <img src="assets/notebook1.png?raw=true" width="49.1%" />
+  <img src="assets/notebook2.png?raw=true" width="48.9%" />
+</p>
+
+## ONNX Export
+
+SAM's lightweight mask decoder can be exported to ONNX format so that it can be run in any environment that supports ONNX runtime, such as in-browser as showcased in the [demo](https://segment-anything.com/demo). Export the model with
+
+```
+python scripts/export_onnx_model.py --checkpoint <path/to/checkpoint> --output <path/to/output>
+```
+
+See the [example notebook](https://github.com/facebookresearch/segment-anything/blob/main/notebooks/onnx_model_example.ipynb) for details on how to combine image preprocessing via SAM's backbone with mask prediction using the ONNX model. It is recommended to use the latest stable version of PyTorch for ONNX export.
+
+## <a name="Models"></a>Model Checkpoints
+
+Three model versions of the model are available with different backbone sizes. These models can be instantiated by running 
+```
+from segment_anything import sam_model_registry
+sam = sam_model_registry["<name>"](checkpoint="<path/to/checkpoint>")
+```
+Click the links below to download the checkpoint for the corresponding model name. The default model in bold can also be instantiated with `build_sam`, as in the examples in [Getting Started](#getting-started).
+
+* **`default` or `vit_h`: [ViT-H SAM model.](https://dl.fbaipublicfiles.com/segment_anything/sam_vit_h_4b8939.pth)**
+* `vit_l`: [ViT-L SAM model.](https://dl.fbaipublicfiles.com/segment_anything/sam_vit_l_0b3195.pth)
+* `vit_b`: [ViT-B SAM model.](https://dl.fbaipublicfiles.com/segment_anything/sam_vit_b_01ec64.pth)
+
+## License
+The model is licensed under the [Apache 2.0 license](LICENSE).
+
+## Contributing
+
+See [contributing](CONTRIBUTING.md) and the [code of conduct](CODE_OF_CONDUCT.md).
+
+## Contributors
+
+The Segment Anything project was made possible with the help of many contributors (alphabetical):
+
+Aaron Adcock, Vaibhav Aggarwal, Morteza Behrooz, Cheng-Yang Fu, Ashley Gabriel, Ahuva Goldstand, Allen Goodman, Sumanth Gurram, Jiabo Hu, Somya Jain, Devansh Kukreja, Robert Kuo, Joshua Lane, Yanghao Li, Lilian Luong, Jitendra Malik, Mallika Malhotra, William Ngan, Omkar Parkhi, Nikhil Raina, Dirk Rowe, Neil Sejoor, Vanessa Stark, Bala Varadarajan, Bram Wasti, Zachary Winstrom

segment_anything/linter.sh

@@ -0,0 +1,32 @@
+#!/bin/bash -e
+# Copyright (c) Facebook, Inc. and its affiliates.
+
+{
+  black --version | grep -E "23\." > /dev/null
+} || {
+  echo "Linter requires 'black==23.*' !"
+  exit 1
+}
+
+ISORT_VERSION=$(isort --version-number)
+if [[ "$ISORT_VERSION" != 5.12* ]]; then
+  echo "Linter requires isort==5.12.0 !"
+  exit 1
+fi
+
+echo "Running isort ..."
+isort . --atomic
+
+echo "Running black ..."
+black -l 100 .
+
+echo "Running flake8 ..."
+if [ -x "$(command -v flake8)" ]; then
+  flake8 .
+else
+  python3 -m flake8 .
+fi
+
+echo "Running mypy..."
+
+mypy --exclude 'setup.py|notebooks' .

segment_anything/notebooks/onnx_model_example.ipynb

@@ -0,0 +1,774 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "901c8ef3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Copyright (c) Meta Platforms, Inc. and affiliates."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1662bb7c",
+   "metadata": {},
+   "source": [
+    "# Produces masks from prompts using an ONNX model"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7fcc21a0",
+   "metadata": {},
+   "source": [
+    "SAM's prompt encoder and mask decoder are very lightweight, which allows for efficient computation of a mask given user input. This notebook shows an example of how to export and use this lightweight component of the model in ONNX format, allowing it to run on a variety of platforms that support an ONNX runtime."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "86daff77",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "\n",
+       "<a target=\"_blank\" href=\"https://colab.research.google.com/github/facebookresearch/segment-anything/blob/main/notebooks/onnx_model_example.ipynb\">\n",
+       "  <img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/>\n",
+       "</a>\n"
+      ],
+      "text/plain": [
+       "<IPython.core.display.HTML object>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "from IPython.display import display, HTML\n",
+    "display(HTML(\n",
+    "\"\"\"\n",
+    "<a target=\"_blank\" href=\"https://colab.research.google.com/github/facebookresearch/segment-anything/blob/main/notebooks/onnx_model_example.ipynb\">\n",
+    "  <img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/>\n",
+    "</a>\n",
+    "\"\"\"\n",
+    "))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "55ae4e00",
+   "metadata": {},
+   "source": [
+    "## Environment Set-up"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "109a5cc2",
+   "metadata": {},
+   "source": [
+    "If running locally using jupyter, first install `segment_anything` in your environment using the [installation instructions](https://github.com/facebookresearch/segment-anything#installation) in the repository. The latest stable versions of PyTorch and ONNX are recommended for this notebook. If running from Google Colab, set `using_collab=True` below and run the cell. In Colab, be sure to select 'GPU' under 'Edit'->'Notebook Settings'->'Hardware accelerator'."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "39b99fc4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "using_colab = False"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "296a69be",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "if using_colab:\n",
+    "    import torch\n",
+    "    import torchvision\n",
+    "    print(\"PyTorch version:\", torch.__version__)\n",
+    "    print(\"Torchvision version:\", torchvision.__version__)\n",
+    "    print(\"CUDA is available:\", torch.cuda.is_available())\n",
+    "    import sys\n",
+    "    !{sys.executable} -m pip install opencv-python matplotlib onnx onnxruntime\n",
+    "    !{sys.executable} -m pip install 'git+https://github.com/facebookresearch/segment-anything.git'\n",
+    "    \n",
+    "    !mkdir images\n",
+    "    !wget -P images https://raw.githubusercontent.com/facebookresearch/segment-anything/main/notebooks/images/truck.jpg\n",
+    "        \n",
+    "    !wget https://dl.fbaipublicfiles.com/segment_anything/sam_vit_h_4b8939.pth"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "dc4a58be",
+   "metadata": {},
+   "source": [
+    "## Set-up"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "42396e8d",
+   "metadata": {},
+   "source": [
+    "Note that this notebook requires both the `onnx` and `onnxruntime` optional dependencies, in addition to `opencv-python` and `matplotlib` for visualization."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2c712610",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import torch\n",
+    "import numpy as np\n",
+    "import cv2\n",
+    "import matplotlib.pyplot as plt\n",
+    "from segment_anything import sam_model_registry, SamPredictor\n",
+    "from segment_anything.utils.onnx import SamOnnxModel\n",
+    "\n",
+    "import onnxruntime\n",
+    "from onnxruntime.quantization import QuantType\n",
+    "from onnxruntime.quantization.quantize import quantize_dynamic"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f29441b9",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def show_mask(mask, ax):\n",
+    "    color = np.array([30/255, 144/255, 255/255, 0.6])\n",
+    "    h, w = mask.shape[-2:]\n",
+    "    mask_image = mask.reshape(h, w, 1) * color.reshape(1, 1, -1)\n",
+    "    ax.imshow(mask_image)\n",
+    "    \n",
+    "def show_points(coords, labels, ax, marker_size=375):\n",
+    "    pos_points = coords[labels==1]\n",
+    "    neg_points = coords[labels==0]\n",
+    "    ax.scatter(pos_points[:, 0], pos_points[:, 1], color='green', marker='*', s=marker_size, edgecolor='white', linewidth=1.25)\n",
+    "    ax.scatter(neg_points[:, 0], neg_points[:, 1], color='red', marker='*', s=marker_size, edgecolor='white', linewidth=1.25)   \n",
+    "    \n",
+    "def show_box(box, ax):\n",
+    "    x0, y0 = box[0], box[1]\n",
+    "    w, h = box[2] - box[0], box[3] - box[1]\n",
+    "    ax.add_patch(plt.Rectangle((x0, y0), w, h, edgecolor='green', facecolor=(0,0,0,0), lw=2))   "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bd0f6b2b",
+   "metadata": {},
+   "source": [
+    "## Export an ONNX model"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1540f719",
+   "metadata": {},
+   "source": [
+    "Set the path below to a SAM model checkpoint, then load the model. This will be needed to both export the model and to calculate embeddings for the model."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "76fc53f4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "checkpoint = \"sam_vit_h_4b8939.pth\"\n",
+    "model_type = \"default\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "11bfc8aa",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "sam = sam_model_registry[model_type](checkpoint=checkpoint)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "450c089c",
+   "metadata": {},
+   "source": [
+    "The script `segment-anything/scripts/export_onnx_model.py` can be used to export the necessary portion of SAM. Alternatively, run the following code to export an ONNX model. If you have already exported a model, set the path below and skip to the next section. Assure that the exported ONNX model aligns with the checkpoint and model type set above. This notebook expects the model was exported with the parameter `return_single_mask=True`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "38a8add8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "onnx_model_path = None  # Set to use an already exported model, then skip to the next section."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7da638ba",
+   "metadata": {
+    "scrolled": false
+   },
+   "outputs": [],
+   "source": [
+    "import warnings\n",
+    "\n",
+    "onnx_model_path = \"sam_onnx_example.onnx\"\n",
+    "\n",
+    "onnx_model = SamOnnxModel(sam, return_single_mask=True)\n",
+    "\n",
+    "dynamic_axes = {\n",
+    "    \"point_coords\": {1: \"num_points\"},\n",
+    "    \"point_labels\": {1: \"num_points\"},\n",
+    "}\n",
+    "\n",
+    "embed_dim = sam.prompt_encoder.embed_dim\n",
+    "embed_size = sam.prompt_encoder.image_embedding_size\n",
+    "mask_input_size = [4 * x for x in embed_size]\n",
+    "dummy_inputs = {\n",
+    "    \"image_embeddings\": torch.randn(1, embed_dim, *embed_size, dtype=torch.float),\n",
+    "    \"point_coords\": torch.randint(low=0, high=1024, size=(1, 5, 2), dtype=torch.float),\n",
+    "    \"point_labels\": torch.randint(low=0, high=4, size=(1, 5), dtype=torch.float),\n",
+    "    \"mask_input\": torch.randn(1, 1, *mask_input_size, dtype=torch.float),\n",
+    "    \"has_mask_input\": torch.tensor([1], dtype=torch.float),\n",
+    "    \"orig_im_size\": torch.tensor([1500, 2250], dtype=torch.float),\n",
+    "}\n",
+    "output_names = [\"masks\", \"iou_predictions\", \"low_res_masks\"]\n",
+    "\n",
+    "with warnings.catch_warnings():\n",
+    "    warnings.filterwarnings(\"ignore\", category=torch.jit.TracerWarning)\n",
+    "    warnings.filterwarnings(\"ignore\", category=UserWarning)\n",
+    "    with open(onnx_model_path, \"wb\") as f:\n",
+    "        torch.onnx.export(\n",
+    "            onnx_model,\n",
+    "            tuple(dummy_inputs.values()),\n",
+    "            f,\n",
+    "            export_params=True,\n",
+    "            verbose=False,\n",
+    "            opset_version=17,\n",
+    "            do_constant_folding=True,\n",
+    "            input_names=list(dummy_inputs.keys()),\n",
+    "            output_names=output_names,\n",
+    "            dynamic_axes=dynamic_axes,\n",
+    "        )    "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c450cf1a",
+   "metadata": {},
+   "source": [
+    "If desired, the model can additionally be quantized and optimized. We find this improves web runtime significantly for negligible change in qualitative performance. Run the next cell to quantize the model, or skip to the next section otherwise."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "235d39fe",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "onnx_model_quantized_path = \"sam_onnx_quantized_example.onnx\"\n",
+    "quantize_dynamic(\n",
+    "    model_input=onnx_model_path,\n",
+    "    model_output=onnx_model_quantized_path,\n",
+    "    optimize_model=True,\n",
+    "    per_channel=False,\n",
+    "    reduce_range=False,\n",
+    "    weight_type=QuantType.QUInt8,\n",
+    ")\n",
+    "onnx_model_path = onnx_model_quantized_path"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "927a928b",
+   "metadata": {},
+   "source": [
+    "## Example Image"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "6be6eb55",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "image = cv2.imread('images/truck.jpg')\n",
+    "image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b7e9a27a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.figure(figsize=(10,10))\n",
+    "plt.imshow(image)\n",
+    "plt.axis('on')\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "027b177b",
+   "metadata": {},
+   "source": [
+    "## Using an ONNX model"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "778d4593",
+   "metadata": {},
+   "source": [
+    "Here as an example, we use `onnxruntime` in python on CPU to execute the ONNX model. However, any platform that supports an ONNX runtime could be used in principle. Launch the runtime session below:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9689b1bf",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "ort_session = onnxruntime.InferenceSession(onnx_model_path)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7708ead6",
+   "metadata": {},
+   "source": [
+    "To use the ONNX model, the image must first be pre-processed using the SAM image encoder. This is a heavier weight process best performed on GPU. SamPredictor can be used as normal, then `.get_image_embedding()` will retreive the intermediate features."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "26e067b4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "sam.to(device='cuda')\n",
+    "predictor = SamPredictor(sam)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7ad3f0d6",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "predictor.set_image(image)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8a6f0f07",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "image_embedding = predictor.get_image_embedding().cpu().numpy()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5e112f33",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "image_embedding.shape"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6337b654",
+   "metadata": {},
+   "source": [
+    "The ONNX model has a different input signature than `SamPredictor.predict`. The following inputs must all be supplied. Note the special cases for both point and mask inputs. All inputs are `np.float32`.\n",
+    "* `image_embeddings`: The image embedding from `predictor.get_image_embedding()`. Has a batch index of length 1.\n",
+    "* `point_coords`: Coordinates of sparse input prompts, corresponding to both point inputs and box inputs. Boxes are encoded using two points, one for the top-left corner and one for the bottom-right corner. *Coordinates must already be transformed to long-side 1024.* Has a batch index of length 1.\n",
+    "* `point_labels`: Labels for the sparse input prompts. 0 is a negative input point, 1 is a positive input point, 2 is a top-left box corner, 3 is a bottom-right box corner, and -1 is a padding point. *If there is no box input, a single padding point with label -1 and coordinates (0.0, 0.0) should be concatenated.*\n",
+    "* `mask_input`: A mask input to the model with shape 1x1x256x256. This must be supplied even if there is no mask input. In this case, it can just be zeros.\n",
+    "* `has_mask_input`: An indicator for the mask input. 1 indicates a mask input, 0 indicates no mask input.\n",
+    "* `orig_im_size`: The size of the input image in (H,W) format, before any transformation. \n",
+    "\n",
+    "Additionally, the ONNX model does not threshold the output mask logits. To obtain a binary mask, threshold at `sam.mask_threshold` (equal to 0.0)."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bf5a9f55",
+   "metadata": {},
+   "source": [
+    "### Example point input"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1c0deef0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "input_point = np.array([[500, 375]])\n",
+    "input_label = np.array([1])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7256394c",
+   "metadata": {},
+   "source": [
+    "Add a batch index, concatenate a padding point, and transform."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "4f69903e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "onnx_coord = np.concatenate([input_point, np.array([[0.0, 0.0]])], axis=0)[None, :, :]\n",
+    "onnx_label = np.concatenate([input_label, np.array([-1])], axis=0)[None, :].astype(np.float32)\n",
+    "\n",
+    "onnx_coord = predictor.transform.apply_coords(onnx_coord, image.shape[:2]).astype(np.float32)\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b188dc53",
+   "metadata": {},
+   "source": [
+    "Create an empty mask input and an indicator for no mask."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5cb52bcf",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "onnx_mask_input = np.zeros((1, 1, 256, 256), dtype=np.float32)\n",
+    "onnx_has_mask_input = np.zeros(1, dtype=np.float32)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a99c2cc5",
+   "metadata": {},
+   "source": [
+    "Package the inputs to run in the onnx model"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b1d7ea11",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "ort_inputs = {\n",
+    "    \"image_embeddings\": image_embedding,\n",
+    "    \"point_coords\": onnx_coord,\n",
+    "    \"point_labels\": onnx_label,\n",
+    "    \"mask_input\": onnx_mask_input,\n",
+    "    \"has_mask_input\": onnx_has_mask_input,\n",
+    "    \"orig_im_size\": np.array(image.shape[:2], dtype=np.float32)\n",
+    "}"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4b6409c9",
+   "metadata": {},
+   "source": [
+    "Predict a mask and threshold it."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "dc4cc082",
+   "metadata": {
+    "scrolled": false
+   },
+   "outputs": [],
+   "source": [
+    "masks, _, low_res_logits = ort_session.run(None, ort_inputs)\n",
+    "masks = masks > predictor.model.mask_threshold"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d778a8fb",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "masks.shape"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "badb1175",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.figure(figsize=(10,10))\n",
+    "plt.imshow(image)\n",
+    "show_mask(masks, plt.gca())\n",
+    "show_points(input_point, input_label, plt.gca())\n",
+    "plt.axis('off')\n",
+    "plt.show() "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1f1d4d15",
+   "metadata": {},
+   "source": [
+    "### Example mask input"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b319da82",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "input_point = np.array([[500, 375], [1125, 625]])\n",
+    "input_label = np.array([1, 1])\n",
+    "\n",
+    "# Use the mask output from the previous run. It is already in the correct form for input to the ONNX model.\n",
+    "onnx_mask_input = low_res_logits"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b1823b37",
+   "metadata": {},
+   "source": [
+    "Transform the points as in the previous example."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8885130f",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "onnx_coord = np.concatenate([input_point, np.array([[0.0, 0.0]])], axis=0)[None, :, :]\n",
+    "onnx_label = np.concatenate([input_label, np.array([-1])], axis=0)[None, :].astype(np.float32)\n",
+    "\n",
+    "onnx_coord = predictor.transform.apply_coords(onnx_coord, image.shape[:2]).astype(np.float32)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "28e47b69",
+   "metadata": {},
+   "source": [
+    "The `has_mask_input` indicator is now 1."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3ab4483a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "onnx_has_mask_input = np.ones(1, dtype=np.float32)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d3781955",
+   "metadata": {},
+   "source": [
+    "Package inputs, then predict and threshold the mask."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "0c1ec096",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "ort_inputs = {\n",
+    "    \"image_embeddings\": image_embedding,\n",
+    "    \"point_coords\": onnx_coord,\n",
+    "    \"point_labels\": onnx_label,\n",
+    "    \"mask_input\": onnx_mask_input,\n",
+    "    \"has_mask_input\": onnx_has_mask_input,\n",
+    "    \"orig_im_size\": np.array(image.shape[:2], dtype=np.float32)\n",
+    "}\n",
+    "\n",
+    "masks, _, _ = ort_session.run(None, ort_inputs)\n",
+    "masks = masks > predictor.model.mask_threshold"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "1e36554b",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.figure(figsize=(10,10))\n",
+    "plt.imshow(image)\n",
+    "show_mask(masks, plt.gca())\n",
+    "show_points(input_point, input_label, plt.gca())\n",
+    "plt.axis('off')\n",
+    "plt.show() "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2ef211d0",
+   "metadata": {},
+   "source": [
+    "### Example box and point input"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "51e58d2e",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "input_box = np.array([425, 600, 700, 875])\n",
+    "input_point = np.array([[575, 750]])\n",
+    "input_label = np.array([0])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6e119dcb",
+   "metadata": {},
+   "source": [
+    "Add a batch index, concatenate a box and point inputs, add the appropriate labels for the box corners, and transform. There is no padding point since the input includes a box input."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "bfbe4911",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "onnx_box_coords = input_box.reshape(2, 2)\n",
+    "onnx_box_labels = np.array([2,3])\n",
+    "\n",
+    "onnx_coord = np.concatenate([input_point, onnx_box_coords], axis=0)[None, :, :]\n",
+    "onnx_label = np.concatenate([input_label, onnx_box_labels], axis=0)[None, :].astype(np.float32)\n",
+    "\n",
+    "onnx_coord = predictor.transform.apply_coords(onnx_coord, image.shape[:2]).astype(np.float32)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "65edabd2",
+   "metadata": {},
+   "source": [
+    "Package inputs, then predict and threshold the mask."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2abfba56",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "onnx_mask_input = np.zeros((1, 1, 256, 256), dtype=np.float32)\n",
+    "onnx_has_mask_input = np.zeros(1, dtype=np.float32)\n",
+    "\n",
+    "ort_inputs = {\n",
+    "    \"image_embeddings\": image_embedding,\n",
+    "    \"point_coords\": onnx_coord,\n",
+    "    \"point_labels\": onnx_label,\n",
+    "    \"mask_input\": onnx_mask_input,\n",
+    "    \"has_mask_input\": onnx_has_mask_input,\n",
+    "    \"orig_im_size\": np.array(image.shape[:2], dtype=np.float32)\n",
+    "}\n",
+    "\n",
+    "masks, _, _ = ort_session.run(None, ort_inputs)\n",
+    "masks = masks > predictor.model.mask_threshold"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "8301bf33",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.figure(figsize=(10, 10))\n",
+    "plt.imshow(image)\n",
+    "show_mask(masks[0], plt.gca())\n",
+    "show_box(input_box, plt.gca())\n",
+    "show_points(input_point, input_label, plt.gca())\n",
+    "plt.axis('off')\n",
+    "plt.show()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

segment_anything/scripts/amg.py

@@ -0,0 +1,238 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+import cv2  # type: ignore
+
+from segment_anything import SamAutomaticMaskGenerator, sam_model_registry
+
+import argparse
+import json
+import os
+from typing import Any, Dict, List
+
+parser = argparse.ArgumentParser(
+    description=(
+        "Runs automatic mask generation on an input image or directory of images, "
+        "and outputs masks as either PNGs or COCO-style RLEs. Requires open-cv, "
+        "as well as pycocotools if saving in RLE format."
+    )
+)
+
+parser.add_argument(
+    "--input",
+    type=str,
+    required=True,
+    help="Path to either a single input image or folder of images.",
+)
+
+parser.add_argument(
+    "--output",
+    type=str,
+    required=True,
+    help=(
+        "Path to the directory where masks will be output. Output will be either a folder "
+        "of PNGs per image or a single json with COCO-style masks."
+    ),
+)
+
+parser.add_argument(
+    "--model-type",
+    type=str,
+    default="default",
+    help="The type of model to load, in ['default', 'vit_l', 'vit_b']",
+)
+
+parser.add_argument(
+    "--checkpoint",
+    type=str,
+    required=True,
+    help="The path to the SAM checkpoint to use for mask generation.",
+)
+
+parser.add_argument("--device", type=str, default="cuda", help="The device to run generation on.")
+
+parser.add_argument(
+    "--convert-to-rle",
+    action="store_true",
+    help=(
+        "Save masks as COCO RLEs in a single json instead of as a folder of PNGs. "
+        "Requires pycocotools."
+    ),
+)
+
+amg_settings = parser.add_argument_group("AMG Settings")
+
+amg_settings.add_argument(
+    "--points-per-side",
+    type=int,
+    default=None,
+    help="Generate masks by sampling a grid over the image with this many points to a side.",
+)
+
+amg_settings.add_argument(
+    "--points-per-batch",
+    type=int,
+    default=None,
+    help="How many input points to process simultaneously in one batch.",
+)
+
+amg_settings.add_argument(
+    "--pred-iou-thresh",
+    type=float,
+    default=None,
+    help="Exclude masks with a predicted score from the model that is lower than this threshold.",
+)
+
+amg_settings.add_argument(
+    "--stability-score-thresh",
+    type=float,
+    default=None,
+    help="Exclude masks with a stability score lower than this threshold.",
+)
+
+amg_settings.add_argument(
+    "--stability-score-offset",
+    type=float,
+    default=None,
+    help="Larger values perturb the mask more when measuring stability score.",
+)
+
+amg_settings.add_argument(
+    "--box-nms-thresh",
+    type=float,
+    default=None,
+    help="The overlap threshold for excluding a duplicate mask.",
+)
+
+amg_settings.add_argument(
+    "--crop-n-layers",
+    type=int,
+    default=None,
+    help=(
+        "If >0, mask generation is run on smaller crops of the image to generate more masks. "
+        "The value sets how many different scales to crop at."
+    ),
+)
+
+amg_settings.add_argument(
+    "--crop-nms-thresh",
+    type=float,
+    default=None,
+    help="The overlap threshold for excluding duplicate masks across different crops.",
+)
+
+amg_settings.add_argument(
+    "--crop-overlap-ratio",
+    type=int,
+    default=None,
+    help="Larger numbers mean image crops will overlap more.",
+)
+
+amg_settings.add_argument(
+    "--crop-n-points-downscale-factor",
+    type=int,
+    default=None,
+    help="The number of points-per-side in each layer of crop is reduced by this factor.",
+)
+
+amg_settings.add_argument(
+    "--min-mask-region-area",
+    type=int,
+    default=None,
+    help=(
+        "Disconnected mask regions or holes with area smaller than this value "
+        "in pixels are removed by postprocessing."
+    ),
+)
+
+
+def write_masks_to_folder(masks: List[Dict[str, Any]], path: str) -> None:
+    header = "id,area,bbox_x0,bbox_y0,bbox_w,bbox_h,point_input_x,point_input_y,predicted_iou,stability_score,crop_box_x0,crop_box_y0,crop_box_w,crop_box_h"  # noqa
+    metadata = [header]
+    for i, mask_data in enumerate(masks):
+        mask = mask_data["segmentation"]
+        filename = f"{i}.png"
+        cv2.imwrite(os.path.join(path, filename), mask * 255)
+        mask_metadata = [
+            str(i),
+            str(mask_data["area"]),
+            *[str(x) for x in mask_data["bbox"]],
+            *[str(x) for x in mask_data["point_coords"][0]],
+            str(mask_data["predicted_iou"]),
+            str(mask_data["stability_score"]),
+            *[str(x) for x in mask_data["crop_box"]],
+        ]
+        row = ",".join(mask_metadata)
+        metadata.append(row)
+    metadata_path = os.path.join(path, "metadata.csv")
+    with open(metadata_path, "w") as f:
+        f.write("\n".join(metadata))
+
+    return
+
+
+def get_amg_kwargs(args):
+    amg_kwargs = {
+        "points_per_side": args.points_per_side,
+        "points_per_batch": args.points_per_batch,
+        "pred_iou_thresh": args.pred_iou_thresh,
+        "stability_score_thresh": args.stability_score_thresh,
+        "stability_score_offset": args.stability_score_offset,
+        "box_nms_thresh": args.box_nms_thresh,
+        "crop_n_layers": args.crop_n_layers,
+        "crop_nms_thresh": args.crop_nms_thresh,
+        "crop_overlap_ratio": args.crop_overlap_ratio,
+        "crop_n_points_downscale_factor": args.crop_n_points_downscale_factor,
+        "min_mask_region_area": args.min_mask_region_area,
+    }
+    amg_kwargs = {k: v for k, v in amg_kwargs.items() if v is not None}
+    return amg_kwargs
+
+
+def main(args: argparse.Namespace) -> None:
+    print("Loading model...")
+    sam = sam_model_registry[args.model_type](checkpoint=args.checkpoint)
+    _ = sam.to(device=args.device)
+    output_mode = "coco_rle" if args.convert_to_rle else "binary_mask"
+    amg_kwargs = get_amg_kwargs(args)
+    generator = SamAutomaticMaskGenerator(sam, output_mode=output_mode, **amg_kwargs)
+
+    if not os.path.isdir(args.input):
+        targets = [args.input]
+    else:
+        targets = [
+            f for f in os.listdir(args.input) if not os.path.isdir(os.path.join(args.input, f))
+        ]
+        targets = [os.path.join(args.input, f) for f in targets]
+
+    os.makedirs(args.output, exist_ok=True)
+
+    for t in targets:
+        print(f"Processing '{t}'...")
+        image = cv2.imread(t)
+        if image is None:
+            print(f"Could not load '{t}' as an image, skipping...")
+            continue
+        image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
+
+        masks = generator.generate(image)
+
+        base = os.path.basename(t)
+        base = os.path.splitext(base)[0]
+        save_base = os.path.join(args.output, base)
+        if output_mode == "binary_mask":
+            os.makedirs(save_base, exist_ok=False)
+            write_masks_to_folder(masks, save_base)
+        else:
+            save_file = save_base + ".json"
+            with open(save_file, "w") as f:
+                json.dump(masks, f)
+    print("Done!")
+
+
+if __name__ == "__main__":
+    args = parser.parse_args()
+    main(args)

segment_anything/scripts/export_onnx_model.py

@@ -0,0 +1,204 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+import torch
+
+from segment_anything import build_sam, build_sam_vit_b, build_sam_vit_l
+from segment_anything.utils.onnx import SamOnnxModel
+
+import argparse
+import warnings
+
+try:
+    import onnxruntime  # type: ignore
+
+    onnxruntime_exists = True
+except ImportError:
+    onnxruntime_exists = False
+
+parser = argparse.ArgumentParser(
+    description="Export the SAM prompt encoder and mask decoder to an ONNX model."
+)
+
+parser.add_argument(
+    "--checkpoint", type=str, required=True, help="The path to the SAM model checkpoint."
+)
+
+parser.add_argument(
+    "--output", type=str, required=True, help="The filename to save the ONNX model to."
+)
+
+parser.add_argument(
+    "--model-type",
+    type=str,
+    default="default",
+    help="In ['default', 'vit_b', 'vit_l']. Which type of SAM model to export.",
+)
+
+parser.add_argument(
+    "--return-single-mask",
+    action="store_true",
+    help=(
+        "If true, the exported ONNX model will only return the best mask, "
+        "instead of returning multiple masks. For high resolution images "
+        "this can improve runtime when upscaling masks is expensive."
+    ),
+)
+
+parser.add_argument(
+    "--opset",
+    type=int,
+    default=17,
+    help="The ONNX opset version to use. Must be >=11",
+)
+
+parser.add_argument(
+    "--quantize-out",
+    type=str,
+    default=None,
+    help=(
+        "If set, will quantize the model and save it with this name. "
+        "Quantization is performed with quantize_dynamic from onnxruntime.quantization.quantize."
+    ),
+)
+
+parser.add_argument(
+    "--gelu-approximate",
+    action="store_true",
+    help=(
+        "Replace GELU operations with approximations using tanh. Useful "
+        "for some runtimes that have slow or unimplemented erf ops, used in GELU."
+    ),
+)
+
+parser.add_argument(
+    "--use-stability-score",
+    action="store_true",
+    help=(
+        "Replaces the model's predicted mask quality score with the stability "
+        "score calculated on the low resolution masks using an offset of 1.0. "
+    ),
+)
+
+parser.add_argument(
+    "--return-extra-metrics",
+    action="store_true",
+    help=(
+        "The model will return five results: (masks, scores, stability_scores, "
+        "areas, low_res_logits) instead of the usual three. This can be "
+        "significantly slower for high resolution outputs."
+    ),
+)
+
+
+def run_export(
+    model_type: str,
+    checkpoint: str,
+    output: str,
+    opset: int,
+    return_single_mask: bool,
+    gelu_approximate: bool = False,
+    use_stability_score: bool = False,
+    return_extra_metrics=False,
+):
+    print("Loading model...")
+    if model_type == "vit_b":
+        sam = build_sam_vit_b(checkpoint)
+    elif model_type == "vit_l":
+        sam = build_sam_vit_l(checkpoint)
+    else:
+        sam = build_sam(checkpoint)
+
+    onnx_model = SamOnnxModel(
+        model=sam,
+        return_single_mask=return_single_mask,
+        use_stability_score=use_stability_score,
+        return_extra_metrics=return_extra_metrics,
+    )
+
+    if gelu_approximate:
+        for n, m in onnx_model.named_modules():
+            if isinstance(m, torch.nn.GELU):
+                m.approximate = "tanh"
+
+    dynamic_axes = {
+        "point_coords": {1: "num_points"},
+        "point_labels": {1: "num_points"},
+    }
+
+    embed_dim = sam.prompt_encoder.embed_dim
+    embed_size = sam.prompt_encoder.image_embedding_size
+    mask_input_size = [4 * x for x in embed_size]
+    dummy_inputs = {
+        "image_embeddings": torch.randn(1, embed_dim, *embed_size, dtype=torch.float),
+        "point_coords": torch.randint(low=0, high=1024, size=(1, 5, 2), dtype=torch.float),
+        "point_labels": torch.randint(low=0, high=4, size=(1, 5), dtype=torch.float),
+        "mask_input": torch.randn(1, 1, *mask_input_size, dtype=torch.float),
+        "has_mask_input": torch.tensor([1], dtype=torch.float),
+        "orig_im_size": torch.tensor([1500, 2250], dtype=torch.float),
+    }
+
+    _ = onnx_model(**dummy_inputs)
+
+    output_names = ["masks", "iou_predictions", "low_res_masks"]
+
+    with warnings.catch_warnings():
+        warnings.filterwarnings("ignore", category=torch.jit.TracerWarning)
+        warnings.filterwarnings("ignore", category=UserWarning)
+        with open(output, "wb") as f:
+            print(f"Exporing onnx model to {output}...")
+            torch.onnx.export(
+                onnx_model,
+                tuple(dummy_inputs.values()),
+                f,
+                export_params=True,
+                verbose=False,
+                opset_version=opset,
+                do_constant_folding=True,
+                input_names=list(dummy_inputs.keys()),
+                output_names=output_names,
+                dynamic_axes=dynamic_axes,
+            )
+
+    if onnxruntime_exists:
+        ort_inputs = {k: to_numpy(v) for k, v in dummy_inputs.items()}
+        ort_session = onnxruntime.InferenceSession(output)
+        _ = ort_session.run(None, ort_inputs)
+        print("Model has successfully been run with ONNXRuntime.")
+
+
+def to_numpy(tensor):
+    return tensor.cpu().numpy()
+
+
+if __name__ == "__main__":
+    args = parser.parse_args()
+    run_export(
+        model_type=args.model_type,
+        checkpoint=args.checkpoint,
+        output=args.output,
+        opset=args.opset,
+        return_single_mask=args.return_single_mask,
+        gelu_approximate=args.gelu_approximate,
+        use_stability_score=args.use_stability_score,
+        return_extra_metrics=args.return_extra_metrics,
+    )
+
+    if args.quantize_out is not None:
+        assert onnxruntime_exists, "onnxruntime is required to quantize the model."
+        from onnxruntime.quantization import QuantType  # type: ignore
+        from onnxruntime.quantization.quantize import quantize_dynamic  # type: ignore
+
+        print(f"Quantizing model and writing to {args.quantize_out}...")
+        quantize_dynamic(
+            model_input=args.output,
+            model_output=args.quantize_out,
+            optimize_model=True,
+            per_channel=False,
+            reduce_range=False,
+            weight_type=QuantType.QUInt8,
+        )
+        print("Done!")

segment_anything/segment_anything.egg-info/PKG-INFO

@@ -0,0 +1,6 @@
+Metadata-Version: 2.1
+Name: segment-anything
+Version: 1.0
+Provides-Extra: all
+Provides-Extra: dev
+License-File: LICENSE

segment_anything/segment_anything.egg-info/SOURCES.txt

@@ -0,0 +1,29 @@
+LICENSE
+README.md
+setup.cfg
+setup.py
+segment_anything/__init__.py
+segment_anything/automatic_mask_generator.py
+segment_anything/build_sam.py
+segment_anything/build_sam_hq.py
+segment_anything/predictor.py
+segment_anything.egg-info/PKG-INFO
+segment_anything.egg-info/SOURCES.txt
+segment_anything.egg-info/dependency_links.txt
+segment_anything.egg-info/requires.txt
+segment_anything.egg-info/top_level.txt
+segment_anything/mobile_encoder/__init__.py
+segment_anything/mobile_encoder/setup_mobile_sam.py
+segment_anything/mobile_encoder/tiny_vit_sam.py
+segment_anything/modeling/__init__.py
+segment_anything/modeling/common.py
+segment_anything/modeling/image_encoder.py
+segment_anything/modeling/mask_decoder.py
+segment_anything/modeling/mask_decoder_hq.py
+segment_anything/modeling/prompt_encoder.py
+segment_anything/modeling/sam.py
+segment_anything/modeling/transformer.py
+segment_anything/utils/__init__.py
+segment_anything/utils/amg.py
+segment_anything/utils/onnx.py
+segment_anything/utils/transforms.py

segment_anything/segment_anything.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

segment_anything/segment_anything.egg-info/requires.txt

@@ -0,0 +1,13 @@
+
+[all]
+matplotlib
+pycocotools
+opencv-python
+onnx
+onnxruntime
+
+[dev]
+flake8
+isort
+black
+mypy

segment_anything/segment_anything.egg-info/top_level.txt

@@ -0,0 +1 @@
+segment_anything

segment_anything/segment_anything/__init__.py

@@ -0,0 +1,23 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+from .build_sam import (
+    build_sam,
+    build_sam_vit_h,
+    build_sam_vit_l,
+    build_sam_vit_b,
+    sam_model_registry,
+)
+from .build_sam_hq import (
+    build_sam_hq,
+    build_sam_hq_vit_h,
+    build_sam_hq_vit_l,
+    build_sam_hq_vit_b,
+    sam_hq_model_registry,
+)
+from .predictor import SamPredictor, SamEncoder, SamDecoder, sam_decode, segment
+from .automatic_mask_generator import SamAutomaticMaskGenerator
+from .mobile_encoder.setup_mobile_sam import setup_model, load_mobile_sam

segment_anything/segment_anything/automatic_mask_generator.py

@@ -0,0 +1,372 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+import numpy as np
+import torch
+from torchvision.ops.boxes import batched_nms, box_area  # type: ignore
+
+from typing import Any, Dict, List, Optional, Tuple
+
+from .modeling import Sam
+from .predictor import SamPredictor
+from .utils.amg import (
+    MaskData,
+    area_from_rle,
+    batch_iterator,
+    batched_mask_to_box,
+    box_xyxy_to_xywh,
+    build_all_layer_point_grids,
+    calculate_stability_score,
+    coco_encode_rle,
+    generate_crop_boxes,
+    is_box_near_crop_edge,
+    mask_to_rle_pytorch,
+    remove_small_regions,
+    rle_to_mask,
+    uncrop_boxes_xyxy,
+    uncrop_masks,
+    uncrop_points,
+)
+
+
+class SamAutomaticMaskGenerator:
+    def __init__(
+        self,
+        model: Sam,
+        points_per_side: Optional[int] = 32,
+        points_per_batch: int = 64,
+        pred_iou_thresh: float = 0.88,
+        stability_score_thresh: float = 0.95,
+        stability_score_offset: float = 1.0,
+        box_nms_thresh: float = 0.7,
+        crop_n_layers: int = 0,
+        crop_nms_thresh: float = 0.7,
+        crop_overlap_ratio: float = 512 / 1500,
+        crop_n_points_downscale_factor: int = 1,
+        point_grids: Optional[List[np.ndarray]] = None,
+        min_mask_region_area: int = 0,
+        output_mode: str = "binary_mask",
+    ) -> None:
+        """
+        Using a SAM model, generates masks for the entire image.
+        Generates a grid of point prompts over the image, then filters
+        low quality and duplicate masks. The default settings are chosen
+        for SAM with a ViT-H backbone.
+
+        Arguments:
+          model (Sam): The SAM model to use for mask prediction.
+          points_per_side (int or None): The number of points to be sampled
+            along one side of the image. The total number of points is
+            points_per_side**2. If None, 'point_grids' must provide explicit
+            point sampling.
+          points_per_batch (int): Sets the number of points run simultaneously
+            by the model. Higher numbers may be faster but use more GPU memory.
+          pred_iou_thresh (float): A filtering threshold in [0,1], using the
+            model's predicted mask quality.
+          stability_score_thresh (float): A filtering threshold in [0,1], using
+            the stability of the mask under changes to the cutoff used to binarize
+            the model's mask predictions.
+          stability_score_offset (float): The amount to shift the cutoff when
+            calculated the stability score.
+          box_nms_thresh (float): The box IoU cutoff used by non-maximal
+            suppression to filter duplicate masks.
+          crops_n_layers (int): If >0, mask prediction will be run again on
+            crops of the image. Sets the number of layers to run, where each
+            layer has 2**i_layer number of image crops.
+          crops_nms_thresh (float): The box IoU cutoff used by non-maximal
+            suppression to filter duplicate masks between different crops.
+          crop_overlap_ratio (float): Sets the degree to which crops overlap.
+            In the first crop layer, crops will overlap by this fraction of
+            the image length. Later layers with more crops scale down this overlap.
+          crop_n_points_downscale_factor (int): The number of points-per-side
+            sampled in layer n is scaled down by crop_n_points_downscale_factor**n.
+          point_grids (list(np.ndarray) or None): A list over explicit grids
+            of points used for sampling, normalized to [0,1]. The nth grid in the
+            list is used in the nth crop layer. Exclusive with points_per_side.
+          min_mask_region_area (int): If >0, postprocessing will be applied
+            to remove disconnected regions and holes in masks with area smaller
+            than min_mask_region_area. Requires opencv.
+          output_mode (str): The form masks are returned in. Can be 'binary_mask',
+            'uncompressed_rle', or 'coco_rle'. 'coco_rle' requires pycocotools.
+            For large resolutions, 'binary_mask' may consume large amounts of
+            memory.
+        """
+
+        assert (points_per_side is None) != (
+            point_grids is None
+        ), "Exactly one of points_per_side or point_grid must be provided."
+        if points_per_side is not None:
+            self.point_grids = build_all_layer_point_grids(
+                points_per_side,
+                crop_n_layers,
+                crop_n_points_downscale_factor,
+            )
+        elif point_grids is not None:
+            self.point_grids = point_grids
+        else:
+            raise ValueError("Can't have both points_per_side and point_grid be None.")
+
+        assert output_mode in [
+            "binary_mask",
+            "uncompressed_rle",
+            "coco_rle",
+        ], f"Unknown output_mode {output_mode}."
+        if output_mode == "coco_rle":
+            from pycocotools import mask as mask_utils  # type: ignore # noqa: F401
+
+        if min_mask_region_area > 0:
+            import cv2  # type: ignore # noqa: F401
+
+        self.predictor = SamPredictor(model)
+        self.points_per_batch = points_per_batch
+        self.pred_iou_thresh = pred_iou_thresh
+        self.stability_score_thresh = stability_score_thresh
+        self.stability_score_offset = stability_score_offset
+        self.box_nms_thresh = box_nms_thresh
+        self.crop_n_layers = crop_n_layers
+        self.crop_nms_thresh = crop_nms_thresh
+        self.crop_overlap_ratio = crop_overlap_ratio
+        self.crop_n_points_downscale_factor = crop_n_points_downscale_factor
+        self.min_mask_region_area = min_mask_region_area
+        self.output_mode = output_mode
+
+    @torch.no_grad()
+    def generate(self, image: np.ndarray) -> List[Dict[str, Any]]:
+        """
+        Generates masks for the given image.
+
+        Arguments:
+          image (np.ndarray): The image to generate masks for, in HWC uint8 format.
+
+        Returns:
+           list(dict(str, any)): A list over records for masks. Each record is
+             a dict containing the following keys:
+               segmentation (dict(str, any) or np.ndarray): The mask. If
+                 output_mode='binary_mask', is an array of shape HW. Otherwise,
+                 is a dictionary containing the RLE.
+               bbox (list(float)): The box around the mask, in XYWH format.
+               area (int): The area in pixels of the mask.
+               predicted_iou (float): The model's own prediction of the mask's
+                 quality. This is filtered by the pred_iou_thresh parameter.
+               point_coords (list(list(float))): The point coordinates input
+                 to the model to generate this mask.
+               stability_score (float): A measure of the mask's quality. This
+                 is filtered on using the stability_score_thresh parameter.
+               crop_box (list(float)): The crop of the image used to generate
+                 the mask, given in XYWH format.
+        """
+
+        # Generate masks
+        mask_data = self._generate_masks(image)
+
+        # Filter small disconnected regions and holes in masks
+        if self.min_mask_region_area > 0:
+            mask_data = self.postprocess_small_regions(
+                mask_data,
+                self.min_mask_region_area,
+                max(self.box_nms_thresh, self.crop_nms_thresh),
+            )
+
+        # Encode masks
+        if self.output_mode == "coco_rle":
+            mask_data["segmentations"] = [coco_encode_rle(rle) for rle in mask_data["rles"]]
+        elif self.output_mode == "binary_mask":
+            mask_data["segmentations"] = [rle_to_mask(rle) for rle in mask_data["rles"]]
+        else:
+            mask_data["segmentations"] = mask_data["rles"]
+
+        # Write mask records
+        curr_anns = []
+        for idx in range(len(mask_data["segmentations"])):
+            ann = {
+                "segmentation": mask_data["segmentations"][idx],
+                "area": area_from_rle(mask_data["rles"][idx]),
+                "bbox": box_xyxy_to_xywh(mask_data["boxes"][idx]).tolist(),
+                "predicted_iou": mask_data["iou_preds"][idx].item(),
+                "point_coords": [mask_data["points"][idx].tolist()],
+                "stability_score": mask_data["stability_score"][idx].item(),
+                "crop_box": box_xyxy_to_xywh(mask_data["crop_boxes"][idx]).tolist(),
+            }
+            curr_anns.append(ann)
+
+        return curr_anns
+
+    def _generate_masks(self, image: np.ndarray) -> MaskData:
+        orig_size = image.shape[:2]
+        crop_boxes, layer_idxs = generate_crop_boxes(
+            orig_size, self.crop_n_layers, self.crop_overlap_ratio
+        )
+
+        # Iterate over image crops
+        data = MaskData()
+        for crop_box, layer_idx in zip(crop_boxes, layer_idxs):
+            crop_data = self._process_crop(image, crop_box, layer_idx, orig_size)
+            data.cat(crop_data)
+
+        # Remove duplicate masks between crops
+        if len(crop_boxes) > 1:
+            # Prefer masks from smaller crops
+            scores = 1 / box_area(data["crop_boxes"])
+            scores = scores.to(data["boxes"].device)
+            keep_by_nms = batched_nms(
+                data["boxes"].float(),
+                scores,
+                torch.zeros(len(data["boxes"])),  # categories
+                iou_threshold=self.crop_nms_thresh,
+            )
+            data.filter(keep_by_nms)
+
+        data.to_numpy()
+        return data
+
+    def _process_crop(
+        self,
+        image: np.ndarray,
+        crop_box: List[int],
+        crop_layer_idx: int,
+        orig_size: Tuple[int, ...],
+    ) -> MaskData:
+        # Crop the image and calculate embeddings
+        x0, y0, x1, y1 = crop_box
+        cropped_im = image[y0:y1, x0:x1, :]
+        cropped_im_size = cropped_im.shape[:2]
+        self.predictor.set_image(cropped_im)
+
+        # Get points for this crop
+        points_scale = np.array(cropped_im_size)[None, ::-1]
+        points_for_image = self.point_grids[crop_layer_idx] * points_scale
+
+        # Generate masks for this crop in batches
+        data = MaskData()
+        for (points,) in batch_iterator(self.points_per_batch, points_for_image):
+            batch_data = self._process_batch(points, cropped_im_size, crop_box, orig_size)
+            data.cat(batch_data)
+            del batch_data
+        self.predictor.reset_image()
+
+        # Remove duplicates within this crop.
+        keep_by_nms = batched_nms(
+            data["boxes"].float(),
+            data["iou_preds"],
+            torch.zeros(len(data["boxes"])),  # categories
+            iou_threshold=self.box_nms_thresh,
+        )
+        data.filter(keep_by_nms)
+
+        # Return to the original image frame
+        data["boxes"] = uncrop_boxes_xyxy(data["boxes"], crop_box)
+        data["points"] = uncrop_points(data["points"], crop_box)
+        data["crop_boxes"] = torch.tensor([crop_box for _ in range(len(data["rles"]))])
+
+        return data
+
+    def _process_batch(
+        self,
+        points: np.ndarray,
+        im_size: Tuple[int, ...],
+        crop_box: List[int],
+        orig_size: Tuple[int, ...],
+    ) -> MaskData:
+        orig_h, orig_w = orig_size
+
+        # Run model on this batch
+        transformed_points = self.predictor.transform.apply_coords(points, im_size)
+        in_points = torch.as_tensor(transformed_points, device=self.predictor.device)
+        in_labels = torch.ones(in_points.shape[0], dtype=torch.int, device=in_points.device)
+        masks, iou_preds, _ = self.predictor.predict_torch(
+            in_points[:, None, :],
+            in_labels[:, None],
+            multimask_output=True,
+            return_logits=True,
+        )
+
+        # Serialize predictions and store in MaskData
+        data = MaskData(
+            masks=masks.flatten(0, 1),
+            iou_preds=iou_preds.flatten(0, 1),
+            points=torch.as_tensor(points.repeat(masks.shape[1], axis=0)),
+        )
+        del masks
+
+        # Filter by predicted IoU
+        if self.pred_iou_thresh > 0.0:
+            keep_mask = data["iou_preds"] > self.pred_iou_thresh
+            data.filter(keep_mask)
+
+        # Calculate stability score
+        data["stability_score"] = calculate_stability_score(
+            data["masks"], self.predictor.model.mask_threshold, self.stability_score_offset
+        )
+        if self.stability_score_thresh > 0.0:
+            keep_mask = data["stability_score"] >= self.stability_score_thresh
+            data.filter(keep_mask)
+
+        # Threshold masks and calculate boxes
+        data["masks"] = data["masks"] > self.predictor.model.mask_threshold
+        data["boxes"] = batched_mask_to_box(data["masks"])
+
+        # Filter boxes that touch crop boundaries
+        keep_mask = ~is_box_near_crop_edge(data["boxes"], crop_box, [0, 0, orig_w, orig_h])
+        if not torch.all(keep_mask):
+            data.filter(keep_mask)
+
+        # Compress to RLE
+        data["masks"] = uncrop_masks(data["masks"], crop_box, orig_h, orig_w)
+        data["rles"] = mask_to_rle_pytorch(data["masks"])
+        del data["masks"]
+
+        return data
+
+    @staticmethod
+    def postprocess_small_regions(
+        mask_data: MaskData, min_area: int, nms_thresh: float
+    ) -> MaskData:
+        """
+        Removes small disconnected regions and holes in masks, then reruns
+        box NMS to remove any new duplicates.
+
+        Edits mask_data in place.
+
+        Requires open-cv as a dependency.
+        """
+        if len(mask_data["rles"]) == 0:
+            return mask_data
+
+        # Filter small disconnected regions and holes
+        new_masks = []
+        scores = []
+        for rle in mask_data["rles"]:
+            mask = rle_to_mask(rle)
+
+            mask, changed = remove_small_regions(mask, min_area, mode="holes")
+            unchanged = not changed
+            mask, changed = remove_small_regions(mask, min_area, mode="islands")
+            unchanged = unchanged and not changed
+
+            new_masks.append(torch.as_tensor(mask).unsqueeze(0))
+            # Give score=0 to changed masks and score=1 to unchanged masks
+            # so NMS will prefer ones that didn't need postprocessing
+            scores.append(float(unchanged))
+
+        # Recalculate boxes and remove any new duplicates
+        masks = torch.cat(new_masks, dim=0)
+        boxes = batched_mask_to_box(masks)
+        keep_by_nms = batched_nms(
+            boxes.float(),
+            torch.as_tensor(scores),
+            torch.zeros(len(boxes)),  # categories
+            iou_threshold=nms_thresh,
+        )
+
+        # Only recalculate RLEs for masks that have changed
+        for i_mask in keep_by_nms:
+            if scores[i_mask] == 0.0:
+                mask_torch = masks[i_mask].unsqueeze(0)
+                mask_data["rles"][i_mask] = mask_to_rle_pytorch(mask_torch)[0]
+                mask_data["boxes"][i_mask] = boxes[i_mask]  # update res directly
+        mask_data.filter(keep_by_nms)
+
+        return mask_data

segment_anything/segment_anything/build_sam.py

@@ -0,0 +1,111 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+import torch
+
+from functools import partial
+
+from .modeling import ImageEncoderViT, MaskDecoder, PromptEncoder, Sam, TwoWayTransformer
+
+def build_sam(checkpoint=None):
+    sam_version = checkpoint.split('.')[0].split('_')[2]
+    if sam_version == 'b':
+        return build_sam_vit_b(checkpoint)
+    elif sam_version == 'l':
+        return build_sam_vit_l(checkpoint)
+    else:
+        return build_sam_vit_h(checkpoint)
+
+def build_sam_vit_h(checkpoint=None):
+    return _build_sam(
+        encoder_embed_dim=1280,
+        encoder_depth=32,
+        encoder_num_heads=16,
+        encoder_global_attn_indexes=[7, 15, 23, 31],
+        checkpoint=checkpoint,
+    )
+
+def build_sam_vit_l(checkpoint=None):
+    return _build_sam(
+        encoder_embed_dim=1024,
+        encoder_depth=24,
+        encoder_num_heads=16,
+        encoder_global_attn_indexes=[5, 11, 17, 23],
+        checkpoint=checkpoint,
+    )
+
+
+def build_sam_vit_b(checkpoint=None):
+    return _build_sam(
+        encoder_embed_dim=768,
+        encoder_depth=12,
+        encoder_num_heads=12,
+        encoder_global_attn_indexes=[2, 5, 8, 11],
+        checkpoint=checkpoint,
+    )
+
+
+sam_model_registry = {
+    "default": build_sam,
+    "vit_h": build_sam,
+    "vit_l": build_sam_vit_l,
+    "vit_b": build_sam_vit_b,
+}
+
+
+def _build_sam(
+    encoder_embed_dim,
+    encoder_depth,
+    encoder_num_heads,
+    encoder_global_attn_indexes,
+    checkpoint=None,
+):
+    prompt_embed_dim = 256
+    image_size = 1024
+    vit_patch_size = 16
+    image_embedding_size = image_size // vit_patch_size
+    sam = Sam(
+        image_encoder=ImageEncoderViT(
+            depth=encoder_depth,
+            embed_dim=encoder_embed_dim,
+            img_size=image_size,
+            mlp_ratio=4,
+            norm_layer=partial(torch.nn.LayerNorm, eps=1e-6),
+            num_heads=encoder_num_heads,
+            patch_size=vit_patch_size,
+            qkv_bias=True,
+            use_rel_pos=True,
+            global_attn_indexes=encoder_global_attn_indexes,
+            window_size=14,
+            out_chans=prompt_embed_dim,
+        ),
+        prompt_encoder=PromptEncoder(
+            embed_dim=prompt_embed_dim,
+            image_embedding_size=(image_embedding_size, image_embedding_size),
+            input_image_size=(image_size, image_size),
+            mask_in_chans=16,
+        ),
+        mask_decoder=MaskDecoder(
+            num_multimask_outputs=3,
+            transformer=TwoWayTransformer(
+                depth=2,
+                embedding_dim=prompt_embed_dim,
+                mlp_dim=2048,
+                num_heads=8,
+            ),
+            transformer_dim=prompt_embed_dim,
+            iou_head_depth=3,
+            iou_head_hidden_dim=256,
+        ),
+        pixel_mean=[123.675, 116.28, 103.53],
+        pixel_std=[58.395, 57.12, 57.375],
+    )
+    sam.eval()
+    if checkpoint is not None:
+        with open(checkpoint, "rb") as f:
+            state_dict = torch.load(f)
+        sam.load_state_dict(state_dict)
+    return sam

segment_anything/segment_anything/build_sam_hq.py

@@ -0,0 +1,117 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+import torch
+
+from functools import partial
+
+from .modeling import ImageEncoderViT, MaskDecoderHQ, PromptEncoder, Sam, TwoWayTransformer
+
+def build_sam_hq(checkpoint=None):
+    sam_version = checkpoint.split('.')[0].split('_')[-1]
+    if sam_version == 'b':
+        return build_sam_hq_vit_b(checkpoint)
+    elif sam_version == 'l':
+        return build_sam_hq_vit_l(checkpoint)
+    else:
+        return build_sam_hq_vit_h(checkpoint)
+
+def build_sam_hq_vit_h(checkpoint=None):
+    return _build_sam(
+        encoder_embed_dim=1280,
+        encoder_depth=32,
+        encoder_num_heads=16,
+        encoder_global_attn_indexes=[7, 15, 23, 31],
+        checkpoint=checkpoint,
+    )
+
+def build_sam_hq_vit_l(checkpoint=None):
+    return _build_sam(
+        encoder_embed_dim=1024,
+        encoder_depth=24,
+        encoder_num_heads=16,
+        encoder_global_attn_indexes=[5, 11, 17, 23],
+        checkpoint=checkpoint,
+    )
+
+
+def build_sam_hq_vit_b(checkpoint=None):
+    return _build_sam(
+        encoder_embed_dim=768,
+        encoder_depth=12,
+        encoder_num_heads=12,
+        encoder_global_attn_indexes=[2, 5, 8, 11],
+        checkpoint=checkpoint,
+    )
+
+
+sam_hq_model_registry = {
+    "default": build_sam_hq_vit_h,
+    "vit_h": build_sam_hq_vit_h,
+    "vit_l": build_sam_hq_vit_l,
+    "vit_b": build_sam_hq_vit_b,
+}
+
+
+def _build_sam(
+    encoder_embed_dim,
+    encoder_depth,
+    encoder_num_heads,
+    encoder_global_attn_indexes,
+    checkpoint=None,
+):
+    prompt_embed_dim = 256
+    image_size = 1024
+    vit_patch_size = 16
+    image_embedding_size = image_size // vit_patch_size
+    sam = Sam(
+        image_encoder=ImageEncoderViT(
+            depth=encoder_depth,
+            embed_dim=encoder_embed_dim,
+            img_size=image_size,
+            mlp_ratio=4,
+            norm_layer=partial(torch.nn.LayerNorm, eps=1e-6),
+            num_heads=encoder_num_heads,
+            patch_size=vit_patch_size,
+            qkv_bias=True,
+            use_rel_pos=True,
+            global_attn_indexes=encoder_global_attn_indexes,
+            window_size=14,
+            out_chans=prompt_embed_dim,
+        ),
+        prompt_encoder=PromptEncoder(
+            embed_dim=prompt_embed_dim,
+            image_embedding_size=(image_embedding_size, image_embedding_size),
+            input_image_size=(image_size, image_size),
+            mask_in_chans=16,
+        ),
+        mask_decoder=MaskDecoderHQ(
+            num_multimask_outputs=3,
+            transformer=TwoWayTransformer(
+                depth=2,
+                embedding_dim=prompt_embed_dim,
+                mlp_dim=2048,
+                num_heads=8,
+            ),
+            transformer_dim=prompt_embed_dim,
+            iou_head_depth=3,
+            iou_head_hidden_dim=256,
+            vit_dim=encoder_embed_dim,
+        ),
+        pixel_mean=[123.675, 116.28, 103.53],
+        pixel_std=[58.395, 57.12, 57.375],
+    )
+    # sam.eval()
+    if checkpoint is not None:
+        with open(checkpoint, "rb") as f:
+            state_dict = torch.load(f)
+        info = sam.load_state_dict(state_dict, strict=False)
+        print(info)
+    for n, p in sam.named_parameters():
+        if 'hf_token' not in n and 'hf_mlp' not in n and 'compress_vit_feat' not in n and 'embedding_encoder' not in n and 'embedding_maskfeature' not in n:
+            p.requires_grad = False
+
+    return sam

segment_anything/segment_anything/mobile_encoder/setup_mobile_sam.py

@@ -0,0 +1,51 @@
+import torch
+from segment_anything.mobile_encoder.tiny_vit_sam import TinyViT
+from segment_anything.modeling import ImageEncoderViT, MaskDecoder, PromptEncoder, Sam, TwoWayTransformer
+def setup_model():
+    prompt_embed_dim = 256
+    image_size = 1024
+    vit_patch_size = 16
+    image_embedding_size = image_size // vit_patch_size
+    mobile_sam = Sam(
+            image_encoder=TinyViT(img_size=1024, in_chans=3, num_classes=1000,
+                embed_dims=[64, 128, 160, 320],
+                depths=[2, 2, 6, 2],
+                num_heads=[2, 4, 5, 10],
+                window_sizes=[7, 7, 14, 7],
+                mlp_ratio=4.,
+                drop_rate=0.,
+                drop_path_rate=0.0,
+                use_checkpoint=False,
+                mbconv_expand_ratio=4.0,
+                local_conv_size=3,
+                layer_lr_decay=0.8
+            ),
+            prompt_encoder=PromptEncoder(
+            embed_dim=prompt_embed_dim,
+            image_embedding_size=(image_embedding_size, image_embedding_size),
+            input_image_size=(image_size, image_size),
+            mask_in_chans=16,
+            ),
+            mask_decoder=MaskDecoder(
+                    num_multimask_outputs=3,
+                    transformer=TwoWayTransformer(
+                    depth=2,
+                    embedding_dim=prompt_embed_dim,
+                    mlp_dim=2048,
+                    num_heads=8,
+                ),
+                transformer_dim=prompt_embed_dim,
+                iou_head_depth=3,
+                iou_head_hidden_dim=256,
+            ),
+            pixel_mean=[123.675, 116.28, 103.53],
+            pixel_std=[58.395, 57.12, 57.375],
+        )
+    return mobile_sam
+def load_mobile_sam(mobile_sam_checkpoint_path,
+                    device='cuda'):
+    checkpoint = torch.load(mobile_sam_checkpoint_path)
+    mobile_sam = setup_model()
+    mobile_sam.load_state_dict(checkpoint, strict=True)
+    mobile_sam.to(device=device)
+    return mobile_sam

segment_anything/segment_anything/mobile_encoder/tiny_vit_sam.py

@@ -0,0 +1,716 @@
+# --------------------------------------------------------
+# TinyViT Model Architecture
+# Copyright (c) 2022 Microsoft
+# Adapted from LeViT and Swin Transformer
+#   LeViT: (https://github.com/facebookresearch/levit)
+#   Swin: (https://github.com/microsoft/swin-transformer)
+# Build the TinyViT Model
+# --------------------------------------------------------
+
+import itertools
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torch.utils.checkpoint as checkpoint
+from timm.models.layers import DropPath as TimmDropPath,\
+    to_2tuple, trunc_normal_
+from timm.models.registry import register_model
+from typing import Tuple
+
+
+class Conv2d_BN(torch.nn.Sequential):
+    def __init__(self, a, b, ks=1, stride=1, pad=0, dilation=1,
+                 groups=1, bn_weight_init=1):
+        super().__init__()
+        self.add_module('c', torch.nn.Conv2d(
+            a, b, ks, stride, pad, dilation, groups, bias=False))
+        bn = torch.nn.BatchNorm2d(b)
+        torch.nn.init.constant_(bn.weight, bn_weight_init)
+        torch.nn.init.constant_(bn.bias, 0)
+        self.add_module('bn', bn)
+
+    @torch.no_grad()
+    def fuse(self):
+        c, bn = self._modules.values()
+        w = bn.weight / (bn.running_var + bn.eps)**0.5
+        w = c.weight * w[:, None, None, None]
+        b = bn.bias - bn.running_mean * bn.weight / \
+            (bn.running_var + bn.eps)**0.5
+        m = torch.nn.Conv2d(w.size(1) * self.c.groups, w.size(
+            0), w.shape[2:], stride=self.c.stride, padding=self.c.padding, dilation=self.c.dilation, groups=self.c.groups)
+        m.weight.data.copy_(w)
+        m.bias.data.copy_(b)
+        return m
+
+
+class DropPath(TimmDropPath):
+    def __init__(self, drop_prob=None):
+        super().__init__(drop_prob=drop_prob)
+        self.drop_prob = drop_prob
+
+    def __repr__(self):
+        msg = super().__repr__()
+        msg += f'(drop_prob={self.drop_prob})'
+        return msg
+
+
+class PatchEmbed(nn.Module):
+    def __init__(self, in_chans, embed_dim, resolution, activation):
+        super().__init__()
+        img_size: Tuple[int, int] = to_2tuple(resolution)
+        self.patches_resolution = (img_size[0] // 4, img_size[1] // 4)
+        self.num_patches = self.patches_resolution[0] * \
+            self.patches_resolution[1]
+        self.in_chans = in_chans
+        self.embed_dim = embed_dim
+        n = embed_dim
+        self.seq = nn.Sequential(
+            Conv2d_BN(in_chans, n // 2, 3, 2, 1),
+            activation(),
+            Conv2d_BN(n // 2, n, 3, 2, 1),
+        )
+
+    def forward(self, x):
+        return self.seq(x)
+
+
+class MBConv(nn.Module):
+    def __init__(self, in_chans, out_chans, expand_ratio,
+                 activation, drop_path):
+        super().__init__()
+        self.in_chans = in_chans
+        self.hidden_chans = int(in_chans * expand_ratio)
+        self.out_chans = out_chans
+
+        self.conv1 = Conv2d_BN(in_chans, self.hidden_chans, ks=1)
+        self.act1 = activation()
+
+        self.conv2 = Conv2d_BN(self.hidden_chans, self.hidden_chans,
+                               ks=3, stride=1, pad=1, groups=self.hidden_chans)
+        self.act2 = activation()
+
+        self.conv3 = Conv2d_BN(
+            self.hidden_chans, out_chans, ks=1, bn_weight_init=0.0)
+        self.act3 = activation()
+
+        self.drop_path = DropPath(
+            drop_path) if drop_path > 0. else nn.Identity()
+
+    def forward(self, x):
+        shortcut = x
+
+        x = self.conv1(x)
+        x = self.act1(x)
+
+        x = self.conv2(x)
+        x = self.act2(x)
+
+        x = self.conv3(x)
+
+        x = self.drop_path(x)
+
+        x += shortcut
+        x = self.act3(x)
+
+        return x
+
+
+class PatchMerging(nn.Module):
+    def __init__(self, input_resolution, dim, out_dim, activation):
+        super().__init__()
+
+        self.input_resolution = input_resolution
+        self.dim = dim
+        self.out_dim = out_dim
+        self.act = activation()
+        self.conv1 = Conv2d_BN(dim, out_dim, 1, 1, 0)
+        stride_c=2
+        if(out_dim==320 or out_dim==448 or out_dim==576):#handongshen  576
+            stride_c=1
+        self.conv2 = Conv2d_BN(out_dim, out_dim, 3, stride_c, 1, groups=out_dim)
+        self.conv3 = Conv2d_BN(out_dim, out_dim, 1, 1, 0)
+
+    def forward(self, x):
+        if x.ndim == 3:
+            H, W = self.input_resolution
+            B = len(x)
+            # (B, C, H, W)
+            x = x.view(B, H, W, -1).permute(0, 3, 1, 2)
+
+        x = self.conv1(x)
+        x = self.act(x)
+
+        x = self.conv2(x)
+        x = self.act(x)
+        x = self.conv3(x)
+        x = x.flatten(2).transpose(1, 2)
+        return x
+
+
+class ConvLayer(nn.Module):
+    def __init__(self, dim, input_resolution, depth,
+                 activation,
+                 drop_path=0., downsample=None, use_checkpoint=False,
+                 out_dim=None,
+                 conv_expand_ratio=4.,
+                 ):
+
+        super().__init__()
+        self.dim = dim
+        self.input_resolution = input_resolution
+        self.depth = depth
+        self.use_checkpoint = use_checkpoint
+
+        # build blocks
+        self.blocks = nn.ModuleList([
+            MBConv(dim, dim, conv_expand_ratio, activation,
+                   drop_path[i] if isinstance(drop_path, list) else drop_path,
+                   )
+            for i in range(depth)])
+
+        # patch merging layer
+        if downsample is not None:
+            self.downsample = downsample(
+                input_resolution, dim=dim, out_dim=out_dim, activation=activation)
+        else:
+            self.downsample = None
+
+    def forward(self, x):
+        for blk in self.blocks:
+            if self.use_checkpoint:
+                x = checkpoint.checkpoint(blk, x)
+            else:
+                x = blk(x)
+        if self.downsample is not None:
+            x = self.downsample(x)
+        return x
+
+
+class Mlp(nn.Module):
+    def __init__(self, in_features, hidden_features=None,
+                 out_features=None, act_layer=nn.GELU, drop=0.):
+        super().__init__()
+        out_features = out_features or in_features
+        hidden_features = hidden_features or in_features
+        self.norm = nn.LayerNorm(in_features)
+        self.fc1 = nn.Linear(in_features, hidden_features)
+        self.fc2 = nn.Linear(hidden_features, out_features)
+        self.act = act_layer()
+        self.drop = nn.Dropout(drop)
+
+    def forward(self, x):
+        x = self.norm(x)
+
+        x = self.fc1(x)
+        x = self.act(x)
+        x = self.drop(x)
+        x = self.fc2(x)
+        x = self.drop(x)
+        return x
+
+
+class Attention(torch.nn.Module):
+    def __init__(self, dim, key_dim, num_heads=8,
+                 attn_ratio=4,
+                 resolution=(14, 14),
+                 ):
+        super().__init__()
+        # (h, w)
+        assert isinstance(resolution, tuple) and len(resolution) == 2
+        self.num_heads = num_heads
+        self.scale = key_dim ** -0.5
+        self.key_dim = key_dim
+        self.nh_kd = nh_kd = key_dim * num_heads
+        self.d = int(attn_ratio * key_dim)
+        self.dh = int(attn_ratio * key_dim) * num_heads
+        self.attn_ratio = attn_ratio
+        h = self.dh + nh_kd * 2
+
+        self.norm = nn.LayerNorm(dim)
+        self.qkv = nn.Linear(dim, h)
+        self.proj = nn.Linear(self.dh, dim)
+
+        points = list(itertools.product(
+            range(resolution[0]), range(resolution[1])))
+        N = len(points)
+        attention_offsets = {}
+        idxs = []
+        for p1 in points:
+            for p2 in points:
+                offset = (abs(p1[0] - p2[0]), abs(p1[1] - p2[1]))
+                if offset not in attention_offsets:
+                    attention_offsets[offset] = len(attention_offsets)
+                idxs.append(attention_offsets[offset])
+        self.attention_biases = torch.nn.Parameter(
+            torch.zeros(num_heads, len(attention_offsets)))
+        self.register_buffer('attention_bias_idxs',
+                             torch.LongTensor(idxs).view(N, N),
+                             persistent=False)
+
+    @torch.no_grad()
+    def train(self, mode=True):
+        super().train(mode)
+        if mode and hasattr(self, 'ab'):
+            del self.ab
+        else:
+            self.ab = self.attention_biases[:, self.attention_bias_idxs]
+
+    def forward(self, x):  # x (B,N,C)
+        B, N, _ = x.shape
+
+        # Normalization
+        x = self.norm(x)
+
+        qkv = self.qkv(x)
+        # (B, N, num_heads, d)
+        q, k, v = qkv.view(B, N, self.num_heads, -
+                           1).split([self.key_dim, self.key_dim, self.d], dim=3)
+        # (B, num_heads, N, d)
+        q = q.permute(0, 2, 1, 3)
+        k = k.permute(0, 2, 1, 3)
+        v = v.permute(0, 2, 1, 3)
+
+        attn = (
+            (q @ k.transpose(-2, -1)) * self.scale
+            +
+            (self.attention_biases[:, self.attention_bias_idxs]
+             if self.training else self.ab)
+        )
+        attn = attn.softmax(dim=-1)
+        x = (attn @ v).transpose(1, 2).reshape(B, N, self.dh)
+        x = self.proj(x)
+        return x
+
+
+class TinyViTBlock(nn.Module):
+    r""" TinyViT Block.
+
+    Args:
+        dim (int): Number of input channels.
+        input_resolution (tuple[int, int]): Input resulotion.
+        num_heads (int): Number of attention heads.
+        window_size (int): Window size.
+        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+        drop (float, optional): Dropout rate. Default: 0.0
+        drop_path (float, optional): Stochastic depth rate. Default: 0.0
+        local_conv_size (int): the kernel size of the convolution between
+                               Attention and MLP. Default: 3
+        activation: the activation function. Default: nn.GELU
+    """
+
+    def __init__(self, dim, input_resolution, num_heads, window_size=7,
+                 mlp_ratio=4., drop=0., drop_path=0.,
+                 local_conv_size=3,
+                 activation=nn.GELU,
+                 ):
+        super().__init__()
+        self.dim = dim
+        self.input_resolution = input_resolution
+        self.num_heads = num_heads
+        assert window_size > 0, 'window_size must be greater than 0'
+        self.window_size = window_size
+        self.mlp_ratio = mlp_ratio
+
+        self.drop_path = DropPath(
+            drop_path) if drop_path > 0. else nn.Identity()
+
+        assert dim % num_heads == 0, 'dim must be divisible by num_heads'
+        head_dim = dim // num_heads
+
+        window_resolution = (window_size, window_size)
+        self.attn = Attention(dim, head_dim, num_heads,
+                              attn_ratio=1, resolution=window_resolution)
+
+        mlp_hidden_dim = int(dim * mlp_ratio)
+        mlp_activation = activation
+        self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim,
+                       act_layer=mlp_activation, drop=drop)
+
+        pad = local_conv_size // 2
+        self.local_conv = Conv2d_BN(
+            dim, dim, ks=local_conv_size, stride=1, pad=pad, groups=dim)
+
+    def forward(self, x):
+        H, W = self.input_resolution
+        B, L, C = x.shape
+        assert L == H * W, "input feature has wrong size"
+        res_x = x
+        if H == self.window_size and W == self.window_size:
+            x = self.attn(x)
+        else:
+            x = x.view(B, H, W, C)
+            pad_b = (self.window_size - H %
+                     self.window_size) % self.window_size
+            pad_r = (self.window_size - W %
+                     self.window_size) % self.window_size
+            padding = pad_b > 0 or pad_r > 0
+
+            if padding:
+                x = F.pad(x, (0, 0, 0, pad_r, 0, pad_b))
+
+            pH, pW = H + pad_b, W + pad_r
+            nH = pH // self.window_size
+            nW = pW // self.window_size
+            # window partition
+            x = x.view(B, nH, self.window_size, nW, self.window_size, C).transpose(2, 3).reshape(
+                B * nH * nW, self.window_size * self.window_size, C)
+            x = self.attn(x)
+            # window reverse
+            x = x.view(B, nH, nW, self.window_size, self.window_size,
+                       C).transpose(2, 3).reshape(B, pH, pW, C)
+
+            if padding:
+                x = x[:, :H, :W].contiguous()
+
+            x = x.view(B, L, C)
+
+        x = res_x + self.drop_path(x)
+
+        x = x.transpose(1, 2).reshape(B, C, H, W)
+        x = self.local_conv(x)
+        x = x.view(B, C, L).transpose(1, 2)
+
+        x = x + self.drop_path(self.mlp(x))
+        return x
+
+    def extra_repr(self) -> str:
+        return f"dim={self.dim}, input_resolution={self.input_resolution}, num_heads={self.num_heads}, " \
+               f"window_size={self.window_size}, mlp_ratio={self.mlp_ratio}"
+
+
+class BasicLayer(nn.Module):
+    """ A basic TinyViT layer for one stage.
+
+    Args:
+        dim (int): Number of input channels.
+        input_resolution (tuple[int]): Input resolution.
+        depth (int): Number of blocks.
+        num_heads (int): Number of attention heads.
+        window_size (int): Local window size.
+        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim.
+        drop (float, optional): Dropout rate. Default: 0.0
+        drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0
+        downsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: None
+        use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False.
+        local_conv_size: the kernel size of the depthwise convolution between attention and MLP. Default: 3
+        activation: the activation function. Default: nn.GELU
+        out_dim: the output dimension of the layer. Default: dim
+    """
+
+    def __init__(self, dim, input_resolution, depth, num_heads, window_size,
+                 mlp_ratio=4., drop=0.,
+                 drop_path=0., downsample=None, use_checkpoint=False,
+                 local_conv_size=3,
+                 activation=nn.GELU,
+                 out_dim=None,
+                 ):
+
+        super().__init__()
+        self.dim = dim
+        self.input_resolution = input_resolution
+        self.depth = depth
+        self.use_checkpoint = use_checkpoint
+
+        # build blocks
+        self.blocks = nn.ModuleList([
+            TinyViTBlock(dim=dim, input_resolution=input_resolution,
+                         num_heads=num_heads, window_size=window_size,
+                         mlp_ratio=mlp_ratio,
+                         drop=drop,
+                         drop_path=drop_path[i] if isinstance(
+                             drop_path, list) else drop_path,
+                         local_conv_size=local_conv_size,
+                         activation=activation,
+                         )
+            for i in range(depth)])
+
+        # patch merging layer
+        if downsample is not None:
+            self.downsample = downsample(
+                input_resolution, dim=dim, out_dim=out_dim, activation=activation)
+        else:
+            self.downsample = None
+
+    def forward(self, x):
+        for blk in self.blocks:
+            if self.use_checkpoint:
+                x = checkpoint.checkpoint(blk, x)
+            else:
+                x = blk(x)
+        if self.downsample is not None:
+            x = self.downsample(x)
+        return x
+
+    def extra_repr(self) -> str:
+        return f"dim={self.dim}, input_resolution={self.input_resolution}, depth={self.depth}"
+
+class LayerNorm2d(nn.Module):
+    def __init__(self, num_channels: int, eps: float = 1e-6) -> None:
+        super().__init__()
+        self.weight = nn.Parameter(torch.ones(num_channels))
+        self.bias = nn.Parameter(torch.zeros(num_channels))
+        self.eps = eps
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        u = x.mean(1, keepdim=True)
+        s = (x - u).pow(2).mean(1, keepdim=True)
+        x = (x - u) / torch.sqrt(s + self.eps)
+        x = self.weight[:, None, None] * x + self.bias[:, None, None]
+        return x
+class TinyViT(nn.Module):
+    def __init__(self, img_size=224, in_chans=3, num_classes=1000,
+                 embed_dims=[96, 192, 384, 768], depths=[2, 2, 6, 2],
+                 num_heads=[3, 6, 12, 24],
+                 window_sizes=[7, 7, 14, 7],
+                 mlp_ratio=4.,
+                 drop_rate=0.,
+                 drop_path_rate=0.1,
+                 use_checkpoint=False,
+                 mbconv_expand_ratio=4.0,
+                 local_conv_size=3,
+                 layer_lr_decay=1.0,
+                 ):
+        super().__init__()
+        self.img_size=img_size
+        self.num_classes = num_classes
+        self.depths = depths
+        self.num_layers = len(depths)
+        self.mlp_ratio = mlp_ratio
+
+        activation = nn.GELU
+
+        self.patch_embed = PatchEmbed(in_chans=in_chans,
+                                      embed_dim=embed_dims[0],
+                                      resolution=img_size,
+                                      activation=activation)
+
+        patches_resolution = self.patch_embed.patches_resolution
+        self.patches_resolution = patches_resolution
+
+        # stochastic depth
+        dpr = [x.item() for x in torch.linspace(0, drop_path_rate,
+                                                sum(depths))]  # stochastic depth decay rule
+
+        # build layers
+        self.layers = nn.ModuleList()
+        for i_layer in range(self.num_layers):
+            kwargs = dict(dim=embed_dims[i_layer],
+                        input_resolution=(patches_resolution[0] // (2 ** (i_layer-1 if i_layer == 3 else i_layer)),
+                                patches_resolution[1] // (2 ** (i_layer-1 if i_layer == 3 else i_layer))),
+                        #   input_resolution=(patches_resolution[0] // (2 ** i_layer),
+                        #                     patches_resolution[1] // (2 ** i_layer)),
+                          depth=depths[i_layer],
+                          drop_path=dpr[sum(depths[:i_layer]):sum(depths[:i_layer + 1])],
+                          downsample=PatchMerging if (
+                              i_layer < self.num_layers - 1) else None,
+                          use_checkpoint=use_checkpoint,
+                          out_dim=embed_dims[min(
+                              i_layer + 1, len(embed_dims) - 1)],
+                          activation=activation,
+                          )
+            if i_layer == 0:
+                layer = ConvLayer(
+                    conv_expand_ratio=mbconv_expand_ratio,
+                    **kwargs,
+                )
+            else:
+                layer = BasicLayer(
+                    num_heads=num_heads[i_layer],
+                    window_size=window_sizes[i_layer],
+                    mlp_ratio=self.mlp_ratio,
+                    drop=drop_rate,
+                    local_conv_size=local_conv_size,
+                    **kwargs)
+            self.layers.append(layer)
+
+        # Classifier head
+        self.norm_head = nn.LayerNorm(embed_dims[-1])
+        self.head = nn.Linear(
+            embed_dims[-1], num_classes) if num_classes > 0 else torch.nn.Identity()
+
+        # init weights
+        self.apply(self._init_weights)
+        self.set_layer_lr_decay(layer_lr_decay)
+        self.neck = nn.Sequential(
+            nn.Conv2d(
+                embed_dims[-1],#handongshen
+                256,
+                kernel_size=1,
+                bias=False,
+            ),
+            LayerNorm2d(256),
+            nn.Conv2d(
+                256,
+                256,
+                kernel_size=3,
+                padding=1,
+                bias=False,
+            ),
+            LayerNorm2d(256),
+        )
+    def set_layer_lr_decay(self, layer_lr_decay):
+        decay_rate = layer_lr_decay
+
+        # layers -> blocks (depth)
+        depth = sum(self.depths)
+        lr_scales = [decay_rate ** (depth - i - 1) for i in range(depth)]
+        print("LR SCALES:", lr_scales)
+
+        def _set_lr_scale(m, scale):
+            for p in m.parameters():
+                p.lr_scale = scale
+
+        self.patch_embed.apply(lambda x: _set_lr_scale(x, lr_scales[0]))
+        i = 0
+        for layer in self.layers:
+            for block in layer.blocks:
+                block.apply(lambda x: _set_lr_scale(x, lr_scales[i]))
+                i += 1
+            if layer.downsample is not None:
+                layer.downsample.apply(
+                    lambda x: _set_lr_scale(x, lr_scales[i - 1]))
+        assert i == depth
+        for m in [self.norm_head, self.head]:
+            m.apply(lambda x: _set_lr_scale(x, lr_scales[-1]))
+
+        for k, p in self.named_parameters():
+            p.param_name = k
+
+        def _check_lr_scale(m):
+            for p in m.parameters():
+                assert hasattr(p, 'lr_scale'), p.param_name
+
+        self.apply(_check_lr_scale)
+
+    def _init_weights(self, m):
+        if isinstance(m, nn.Linear):
+            trunc_normal_(m.weight, std=.02)
+            if isinstance(m, nn.Linear) and m.bias is not None:
+                nn.init.constant_(m.bias, 0)
+        elif isinstance(m, nn.LayerNorm):
+            nn.init.constant_(m.bias, 0)
+            nn.init.constant_(m.weight, 1.0)
+
+    @torch.jit.ignore
+    def no_weight_decay_keywords(self):
+        return {'attention_biases'}
+
+    def forward_features(self, x):
+        # x: (N, C, H, W)
+        x = self.patch_embed(x)
+
+        x = self.layers[0](x)
+        start_i = 1
+
+        for i in range(start_i, len(self.layers)):
+            layer = self.layers[i]
+            x = layer(x)
+        B,_,C=x.size()
+        x = x.view(B, 64, 64, C)
+        x=x.permute(0, 3, 1, 2)
+        x=self.neck(x)
+        return x
+
+    def forward(self, x):
+        x = self.forward_features(x)
+        #x = self.norm_head(x)
+        #x = self.head(x)
+        return x
+
+
+_checkpoint_url_format = \
+    'https://github.com/wkcn/TinyViT-model-zoo/releases/download/checkpoints/{}.pth'
+_provided_checkpoints = {
+    'tiny_vit_5m_224': 'tiny_vit_5m_22kto1k_distill',
+    'tiny_vit_11m_224': 'tiny_vit_11m_22kto1k_distill',
+    'tiny_vit_21m_224': 'tiny_vit_21m_22kto1k_distill',
+    'tiny_vit_21m_384': 'tiny_vit_21m_22kto1k_384_distill',
+    'tiny_vit_21m_512': 'tiny_vit_21m_22kto1k_512_distill',
+}
+
+
+def register_tiny_vit_model(fn):
+    '''Register a TinyViT model
+    It is a wrapper of `register_model` with loading the pretrained checkpoint.
+    '''
+    def fn_wrapper(pretrained=False, **kwargs):
+        model = fn()
+        if pretrained:
+            model_name = fn.__name__
+            assert model_name in _provided_checkpoints, \
+                f'Sorry that the checkpoint `{model_name}` is not provided yet.'
+            url = _checkpoint_url_format.format(
+                _provided_checkpoints[model_name])
+            checkpoint = torch.hub.load_state_dict_from_url(
+                url=url,
+                map_location='cpu', check_hash=False,
+            )
+            model.load_state_dict(checkpoint['model'])
+
+        return model
+
+    # rename the name of fn_wrapper
+    fn_wrapper.__name__ = fn.__name__
+    return register_model(fn_wrapper)
+
+
+@register_tiny_vit_model
+def tiny_vit_5m_224(pretrained=False, num_classes=1000, drop_path_rate=0.0):
+    return TinyViT(
+        num_classes=num_classes,
+        embed_dims=[64, 128, 160, 320],
+        depths=[2, 2, 6, 2],
+        num_heads=[2, 4, 5, 10],
+        window_sizes=[7, 7, 14, 7],
+        drop_path_rate=drop_path_rate,
+    )
+
+
+@register_tiny_vit_model
+def tiny_vit_11m_224(pretrained=False, num_classes=1000, drop_path_rate=0.1):
+    return TinyViT(
+        num_classes=num_classes,
+        embed_dims=[64, 128, 256, 448],
+        depths=[2, 2, 6, 2],
+        num_heads=[2, 4, 8, 14],
+        window_sizes=[7, 7, 14, 7],
+        drop_path_rate=drop_path_rate,
+    )
+
+
+@register_tiny_vit_model
+def tiny_vit_21m_224(pretrained=False, num_classes=1000, drop_path_rate=0.2):
+    return TinyViT(
+        num_classes=num_classes,
+        embed_dims=[96, 192, 384, 576],
+        depths=[2, 2, 6, 2],
+        num_heads=[3, 6, 12, 18],
+        window_sizes=[7, 7, 14, 7],
+        drop_path_rate=drop_path_rate,
+    )
+
+
+@register_tiny_vit_model
+def tiny_vit_21m_384(pretrained=False, num_classes=1000, drop_path_rate=0.1):
+    return TinyViT(
+        img_size=384,
+        num_classes=num_classes,
+        embed_dims=[96, 192, 384, 576],
+        depths=[2, 2, 6, 2],
+        num_heads=[3, 6, 12, 18],
+        window_sizes=[12, 12, 24, 12],
+        drop_path_rate=drop_path_rate,
+    )
+
+
+@register_tiny_vit_model
+def tiny_vit_21m_512(pretrained=False, num_classes=1000, drop_path_rate=0.1):
+    return TinyViT(
+        img_size=512,
+        num_classes=num_classes,
+        embed_dims=[96, 192, 384, 576],
+        depths=[2, 2, 6, 2],
+        num_heads=[3, 6, 12, 18],
+        window_sizes=[16, 16, 32, 16],
+        drop_path_rate=drop_path_rate,
+    )

segment_anything/segment_anything/modeling/__init__.py

@@ -0,0 +1,12 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+
+# This source code is licensed under the license found in the
+# LICENSE file in the root directory of this source tree.
+
+from .sam import Sam
+from .image_encoder import ImageEncoderViT
+from .mask_decoder_hq import MaskDecoderHQ
+from .mask_decoder import MaskDecoder
+from .prompt_encoder import PromptEncoder
+from .transformer import TwoWayTransformer