{
    "info": {
        "author": "Jonas Rauber & Wieland Brendel",
        "author_email": "opensource@bethgelab.org",
        "bugtrack_url": null,
        "classifiers": [
            "Development Status :: 3 - Alpha",
            "Intended Audience :: Developers",
            "Intended Audience :: Science/Research",
            "License :: OSI Approved :: MIT License",
            "Programming Language :: Python :: 2.7",
            "Programming Language :: Python :: 3",
            "Programming Language :: Python :: 3.5",
            "Programming Language :: Python :: 3.6",
            "Topic :: Scientific/Engineering :: Artificial Intelligence"
        ],
        "description": ".. image:: https://travis-ci.org/bethgelab/robust-vision-benchmark.svg?branch=master\n    :target: https://travis-ci.org/bethgelab/robust-vision-benchmark\n\n=======================\nRobust Vision Benchmark\n=======================\n\nThis Python package provides utility functions to create submissions for the `Robust Vision Benchmark <https://robust.vision/benchmark>`__ and scripts to automatically test and upload them. You might also want to have a look at `Foolbox <https://github.com/bethgelab/foolbox>`__, our python toolbox to benchmark the robustness of machine learning models using a large set of adversarial attacks.\n\nInstallation\n------------\n\nWe test using Python 2.7, 3.5 and 3.6. Other Python versions might work as well, but we recommend using Python 3.5 or newer.\n\n.. code-block:: bash\n\n   pip install robust-vision-benchmark\n\nSubmitting a model\n------------------\n\nA model submission consists of a Dockerfile, a script (e.g. Python) and data (e.g. network weights).\n\nPython script\n^^^^^^^^^^^^^\n\nCreate a Python script that turns your model into a Foolbox model using one of our wrappers for TensorFlow, PyTorch, Theano, Keras, Lasagne, MXNet and starts the `model_server`.\n\n.. code-block:: python\n\n   from robust_vision_benchmark import model_server\n\n   # create your model\n   # ...\n\n   # turn it into a Foolbox model\n   model = foolbox.models.SomeModel(...)\n\n   # start the server\n   mnist_model_server(model)\n   cifar_model_server(model, channel_order='RGB or BGR')\n   imagenet_model_server(model, channel_order='RGB or BGR', image_size=224)\n\n   # For CIFAR and Imagenet, the channel_order must be set to either 'RGB' or 'BGR'.\n   # For ImageNet, the image_size must be set to an integer (usually 224 vor VGG-like networks and 299 for inception-like networks).\n\nDockerfile\n^^^^^^^^^^\n\nCreate a Dockerfile that installs all dependencies and starts the script.\n\n.. code-block:: Dockerfile\n\n   FROM nvidia/cuda:8.0-cudnn5-devel-ubuntu16.04\n\n   RUN apt-get update ... && install ...\n\n   RUN pip3 install --no-cache-dir robust-vision-benchmark\n\n   ...\n\n   COPY main.py main.py\n\n   CMD [\"python3\", \"./main.py\"]\n\nFor compatibility with our backend, please use a derivative of the `nvidia/cuda:8.0` image (e.g. `nvidia/cuda8.0-cudnn5-devel-ubuntu16.04`) as the base image or contact us if you have special requirements.\n\nTest the submission\n^^^^^^^^^^^^^^^^^^^\n\nPut the Dockerfile, script, data and other required files into a folder, e.g. *model* and run the following in your shell:\n\n.. code-block:: bash\n\n   rvb-test-model model/\n\nYou can find an example in *examples/model/*.\n\nUpload the submission\n^^^^^^^^^^^^^^^^^^^^^\n\nOnce your model is ready for submission, upload it:\n\n.. code-block:: bash\n\n   rvb-upload model/\n\nSubmitting\n^^^^^^^^^^\n\nGo to https://robust.vision/benchmark/participate and put the URL returned by the upload script into the `Submission URL`.\n\nSubmitting an attack\n--------------------\n\nAn attack submission consists of a Dockerfile and a script (e.g. Python).\n\nPython script\n^^^^^^^^^^^^^\n\nCreate a Python script that implements your attack and starts the `attack_server`.\n\n.. code-block:: python\n\n   from robust_vision_benchmark import attack_server\n\n   # implement your attack\n   def attack(a):\n       # ...\n\n   # start the server\n   attack_server(attack)\n\nDockerfile\n^^^^^^^^^^\n\nCreate a Dockerfile that installs all dependencies and starts the script.\n\n.. code-block:: Dockerfile\n\n   FROM python:3.6\n\n   RUN pip3 install --no-cache-dir robust-vision-benchmark\n\n   ...\n\n   COPY main.py main.py\n\n   CMD [\"python3\", \"./main.py\"]\n\nTest the submission\n^^^^^^^^^^^^^^^^^^^\n\nPut the Dockerfile, script and other required files into a folder, e.g. *attack* and run the following in your shell:\n\n.. code-block:: bash\n\n   rvb-test-attack attack/\n\nYou can find an example in *examples/attack/*.\n\nUpload the submission\n^^^^^^^^^^^^^^^^^^^^^\n\nOnce your attack is ready for submission, upload it:\n\n.. code-block:: bash\n\n   rvb-upload attack/\n\nSubmitting\n^^^^^^^^^^\n\nGo to https://robust.vision/benchmark/participate and put the URL returned by the upload script into the `Submission URL`.\n\nAuthors\n-------\n\n* `Jonas Rauber <https://github.com/jonasrauber>`_\n* `Wieland Brendel <https://github.com/wielandbrendel>`_",
        "description_content_type": null,
        "docs_url": null,
        "download_url": "",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/bethgelab/robust-vision-benchmark",
        "keywords": "",
        "license": "MIT",
        "maintainer": "",
        "maintainer_email": "",
        "name": "robust-vision-benchmark",
        "package_url": "https://pypi.org/project/robust-vision-benchmark/",
        "platform": "",
        "project_url": "https://pypi.org/project/robust-vision-benchmark/",
        "project_urls": {
            "Homepage": "https://github.com/bethgelab/robust-vision-benchmark"
        },
        "release_url": "https://pypi.org/project/robust-vision-benchmark/0.9.1/",
        "requires_dist": null,
        "requires_python": "",
        "summary": "An HTTP server and client that provides access to Foolbox models and attacks.",
        "version": "0.9.1"
    },
    "last_serial": 3262446,
    "releases": {
        "0.8.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "c2d6c23a13e05831adc1da0db3d25a68",
                    "sha256": "d37d789f1b535a51ac68115e65211ec4c39866565a7fa11a38c75238954028c6"
                },
                "downloads": -1,
                "filename": "robust_vision_benchmark-0.8.0.tar.gz",
                "has_sig": false,
                "md5_digest": "c2d6c23a13e05831adc1da0db3d25a68",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 12318332,
                "upload_time": "2017-08-11T00:06:02",
                "url": "https://files.pythonhosted.org/packages/b1/58/bdc49ae2aff5d79670442fc9a5cd72d4a128a1d95b02e958c8ab0ef23201/robust_vision_benchmark-0.8.0.tar.gz"
            }
        ],
        "0.8.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "3993b64ce6c7decdadafad96ee0cb9d0",
                    "sha256": "ca1a5fee412dd26c50f35371870a8e8bf4096291911abfb6e212d49625184335"
                },
                "downloads": -1,
                "filename": "robust_vision_benchmark-0.8.1.tar.gz",
                "has_sig": false,
                "md5_digest": "3993b64ce6c7decdadafad96ee0cb9d0",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 12318385,
                "upload_time": "2017-08-11T02:59:57",
                "url": "https://files.pythonhosted.org/packages/65/82/0c9a65f18a00c039a5d4304a64f0434c1cbe2de91127049a0b27371c4cf4/robust_vision_benchmark-0.8.1.tar.gz"
            }
        ],
        "0.8.2": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "2470b619a2a5f5680570ea0675e90550",
                    "sha256": "bb1026b3585563ccb17fffbe9d926fee6f20d37a5b84a8fdc3831698b78cd59b"
                },
                "downloads": -1,
                "filename": "robust_vision_benchmark-0.8.2.tar.gz",
                "has_sig": false,
                "md5_digest": "2470b619a2a5f5680570ea0675e90550",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 12319059,
                "upload_time": "2017-09-05T08:50:15",
                "url": "https://files.pythonhosted.org/packages/55/54/5b96f4db2e5703a07bd75239fc15fa7a8f42b69e69969412294b0914a9c7/robust_vision_benchmark-0.8.2.tar.gz"
            }
        ],
        "0.9.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "5328afb11049ded71d7a53a3535b30cc",
                    "sha256": "725f35421b151f795de9d6517f2b0883e7b587955f9b7453811c25cf5c3f430d"
                },
                "downloads": -1,
                "filename": "robust_vision_benchmark-0.9.0.tar.gz",
                "has_sig": false,
                "md5_digest": "5328afb11049ded71d7a53a3535b30cc",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 12319146,
                "upload_time": "2017-09-05T09:49:27",
                "url": "https://files.pythonhosted.org/packages/50/99/0fa5b8c8c01c80cab3d3d7bc88372f55af61056db105e9e30044e1135b62/robust_vision_benchmark-0.9.0.tar.gz"
            }
        ],
        "0.9.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "434aa0038d77206141074cebf26f09d2",
                    "sha256": "d598c5d96a56e6fbda8ac00eec7762dc0864620b246a2ad0e9df3c0f0271fb5a"
                },
                "downloads": -1,
                "filename": "robust_vision_benchmark-0.9.1.tar.gz",
                "has_sig": false,
                "md5_digest": "434aa0038d77206141074cebf26f09d2",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 12319113,
                "upload_time": "2017-10-19T09:32:53",
                "url": "https://files.pythonhosted.org/packages/e7/89/8c93ff7b257c50fa79c065d63a724fb24c40bc62304c7e5557137f7d9b93/robust_vision_benchmark-0.9.1.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "434aa0038d77206141074cebf26f09d2",
                "sha256": "d598c5d96a56e6fbda8ac00eec7762dc0864620b246a2ad0e9df3c0f0271fb5a"
            },
            "downloads": -1,
            "filename": "robust_vision_benchmark-0.9.1.tar.gz",
            "has_sig": false,
            "md5_digest": "434aa0038d77206141074cebf26f09d2",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 12319113,
            "upload_time": "2017-10-19T09:32:53",
            "url": "https://files.pythonhosted.org/packages/e7/89/8c93ff7b257c50fa79c065d63a724fb24c40bc62304c7e5557137f7d9b93/robust_vision_benchmark-0.9.1.tar.gz"
        }
    ]
}