{
    "info": {
        "author": "Robin Br\u00fcgger",
        "author_email": "brueggerrobin+revtorch@gmail.com",
        "bugtrack_url": null,
        "classifiers": [
            "Development Status :: 4 - Beta",
            "Intended Audience :: Science/Research",
            "License :: OSI Approved :: BSD License",
            "Programming Language :: Python :: 3",
            "Programming Language :: Python :: 3.4",
            "Programming Language :: Python :: 3.5",
            "Programming Language :: Python :: 3.6",
            "Programming Language :: Python :: 3.7",
            "Programming Language :: Python :: 3.8",
            "Topic :: Scientific/Engineering :: Artificial Intelligence"
        ],
        "description": "# RevTorch\nFramework for creating (partially) reversible neural networks with PyTorch\n\nRevTorch is introduced and explained in our paper [A Partially Reversible U-Net for Memory-Efficient Volumetric Image Segmentation](https://arxiv.org/abs/1906.06148),\nwhich was accepted for presentation at [MICCAI 2019](https://www.miccai2019.org/). \n\nIf you find this code helpful in your research please cite the following paper:\n```\n@article{PartiallyRevUnet2019Bruegger,\n         author={Br{\\\"u}gger, Robin and Baumgartner, Christian F.\n         and Konukoglu, Ender},\n         title={A Partially Reversible U-Net for Memory-Efficient Volumetric Image Segmentation},\n         journal={arXiv:1906.06148},\n         year={2019},\n```\n\n## Installation\nUse pip to install RevTorch:\n```sh\n$ pip install revtorch\n```\nRevTorch requires PyTorch. However, PyTorch is not included in the dependencies since the required PyTorch version is dependent on your system. Please install PyTorch following the instructions on the [PyTorch website](https://pytorch.org/).\n\n## Usage\nThis example shows how to use the RevTorch framework.\n```python\nimport torch\nimport torchvision\nimport torchvision.transforms as transforms\nimport torch.nn as nn\nimport torch.nn.functional as F\nimport torch.optim as optim\nimport revtorch as rv\n\ndef train():\n    trainset = torchvision.datasets.CIFAR10(root=\"./data\", train=True, download=True, transform=transforms.ToTensor())\n    trainloader = torch.utils.data.DataLoader(trainset, batch_size=4, shuffle=True)\n\n    net = PartiallyReversibleNet()\n    criterion = nn.CrossEntropyLoss()\n    optimizer = optim.Adam(net.parameters())\n\n    for epoch in range(2):\n\n        running_loss = 0.0\n        for i, data in enumerate(trainloader, 0):\n            inputs, labels = data\n            optimizer.zero_grad()\n            outputs = net(inputs)\n            loss = criterion(outputs, labels)\n            loss.backward()\n            optimizer.step()\n\n            #logging stuff\n            running_loss += loss.item()\n            LOG_INTERVAL = 200\n            if i % LOG_INTERVAL == (LOG_INTERVAL-1):  # print every 2000 mini-batches\n                print('[%d, %5d] loss: %.3f' % (epoch + 1, i + 1, running_loss / LOG_INTERVAL))\n                running_loss = 0.0\n\nclass PartiallyReversibleNet(nn.Module):\n    def __init__(self):\n        super(PartiallyReversibleNet, self).__init__()\n\n        #initial non-reversible convolution to get to 32 channels\n        self.conv1 = nn.Conv2d(3, 32, 3)\n\n        #construct reversible sequencce with 4 reversible blocks\n        blocks = []\n        for i in range(4):\n\n            #f and g must both be a nn.Module whos output has the same shape as its input\n            f_func = nn.Sequential(nn.ReLU(), nn.Conv2d(16, 16, 3, padding=1))\n            g_func = nn.Sequential(nn.ReLU(), nn.Conv2d(16, 16, 3, padding=1))\n\n            #we construct a reversible block with our F and G functions\n            blocks.append(rv.ReversibleBlock(f_func, g_func))\n\n        #pack all reversible blocks into a reversible sequence\n        self.sequence = rv.ReversibleSequence(nn.ModuleList(blocks))\n\n        #non-reversible convolution to get to 10 channels (one for each label)\n        self.conv2 = nn.Conv2d(32, 10, 3)\n\n    def forward(self, x):\n        x = self.conv1(x)\n\n        #the reversible sequence can be used like any other nn.Module. Memory-saving backpropagation is used automatically\n        x = self.sequence(x)\n\n        x = self.conv2(F.relu(x))\n        x = F.avg_pool2d(x, (x.shape[2], x.shape[3]))\n        x = x.view(x.shape[0], x.shape[1])\n        return x\n\nif __name__ == \"__main__\":\n    train()\n```\n\n## Python version\nTested with Python 3.6 and PyTorch 1.1.0. Should work with any version of Python 3.",
        "description_content_type": "text/markdown",
        "docs_url": null,
        "download_url": "https://github.com/RobinBruegger/RevTorch/archive/v0.2.0.tar.gz",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/RobinBruegger/RevTorch",
        "keywords": "reversbile neural network",
        "license": "bsd-3-clause",
        "maintainer": "",
        "maintainer_email": "",
        "name": "revtorch",
        "package_url": "https://pypi.org/project/revtorch/",
        "platform": "",
        "project_url": "https://pypi.org/project/revtorch/",
        "project_urls": {
            "Download": "https://github.com/RobinBruegger/RevTorch/archive/v0.2.0.tar.gz",
            "Homepage": "https://github.com/RobinBruegger/RevTorch"
        },
        "release_url": "https://pypi.org/project/revtorch/0.2.0/",
        "requires_dist": null,
        "requires_python": "",
        "summary": "Framework for creating (partially) reversible neural networks with PyTorch",
        "version": "0.2.0"
    },
    "last_serial": 5832381,
    "releases": {
        "0.1.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "883389f9f732fb16808561a6404452d0",
                    "sha256": "af9307d1250286c7c16a23c19f711af33b3dccc2cd98f1bc7b2e91cacfcd2695"
                },
                "downloads": -1,
                "filename": "revtorch-0.1.0.tar.gz",
                "has_sig": false,
                "md5_digest": "883389f9f732fb16808561a6404452d0",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 2819,
                "upload_time": "2019-06-13T11:41:28",
                "url": "https://files.pythonhosted.org/packages/df/e3/28327129901ac65131ad8a948060a06642f0645ac52a1b25a1f7a0540c04/revtorch-0.1.0.tar.gz"
            }
        ],
        "0.1.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "f192076fb02e8261cb720ce340ce987a",
                    "sha256": "a8bf17a05b8a4c7a3d0e559305ce3080d6fbdd6ab463764f03ad970541fa2ad6"
                },
                "downloads": -1,
                "filename": "revtorch-0.1.1.tar.gz",
                "has_sig": false,
                "md5_digest": "f192076fb02e8261cb720ce340ce987a",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 4946,
                "upload_time": "2019-06-18T15:56:07",
                "url": "https://files.pythonhosted.org/packages/e0/f0/10869ad425c8de0bee0ebd254d1ec75af69e9e1c7c97a2b4e828deede789/revtorch-0.1.1.tar.gz"
            }
        ],
        "0.2.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "0681cc6f6891e1572ccaaa86b44b477d",
                    "sha256": "107b72f9dc3a8a581bb57e5909fea414bda2f738e916ffea5a0338c1cf3c9862"
                },
                "downloads": -1,
                "filename": "revtorch-0.2.0.tar.gz",
                "has_sig": false,
                "md5_digest": "0681cc6f6891e1572ccaaa86b44b477d",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 4934,
                "upload_time": "2019-09-15T15:50:56",
                "url": "https://files.pythonhosted.org/packages/ee/af/6b44c2def528b4d80e1f5719ef5954422e2efe4efeccc569d7075292054d/revtorch-0.2.0.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "0681cc6f6891e1572ccaaa86b44b477d",
                "sha256": "107b72f9dc3a8a581bb57e5909fea414bda2f738e916ffea5a0338c1cf3c9862"
            },
            "downloads": -1,
            "filename": "revtorch-0.2.0.tar.gz",
            "has_sig": false,
            "md5_digest": "0681cc6f6891e1572ccaaa86b44b477d",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 4934,
            "upload_time": "2019-09-15T15:50:56",
            "url": "https://files.pythonhosted.org/packages/ee/af/6b44c2def528b4d80e1f5719ef5954422e2efe4efeccc569d7075292054d/revtorch-0.2.0.tar.gz"
        }
    ]
}