{
    "info": {
        "author": "Giuseppe Cuccu",
        "author_email": "giuseppe.cuccu@gmail.com",
        "bugtrack_url": null,
        "classifiers": [
            "Development Status :: 4 - Beta",
            "Intended Audience :: Developers",
            "License :: OSI Approved :: MIT License",
            "Programming Language :: Python :: 3.6",
            "Programming Language :: Python :: 3.7",
            "Programming Language :: Python :: 3.8",
            "Programming Language :: Python :: 3 :: Only",
            "Topic :: Scientific/Engineering :: Artificial Intelligence",
            "Topic :: Software Development :: Build Tools"
        ],
        "description": "# tinynet\n\nA tiny neural network library\n\n## No training\n\nThis library provides no training algorithm. You can easily set up a neuroevolution framework by including any black-box search algorithm, from [RWG](https://www.bioinf.jku.at/publications/older/ch9.pdf) (see example below) to [CMA-ES](https://github.com/CMA-ES/pycma).\n\n## Installation\n\n`pip install tinynet`\n\n## Usage\n\n```python\nfrom tinynet import RNN1L\nimport numpy as np\nnet_struct = [3, 5, 2]\nnet = RNN(net_struct) # try also FFNN\nnet.set_weights(np.random.randn(net.nweights))\nout = net.activate(np.zeros(net.ninputs))\nassert len(out) == net.noutputs\nassert len(net.state) == net.ninputs + 1 + net.noutputs # input, bias, recursion\n```\n\n## Neuroevolution application on the OpenAI Gym\n\nCheck out [this GitHub gist](https://gist.github.com/giuse/3d16c947259173d571cf82e28a2f7a7e) to run the Bipedal Walker using pre-trained weights.\n\nThe example below tackles the CartPole from scratch using RWG.\n\n\n```python\nimport numpy as np\nimport tinynet\nimport gym # just `pip install gym`\n\n# Environment setup\nenv = gym.make(\"CartPole-v1\")\n# env = gym.wrappers.Monitor(env, 'video', force = True) # Uncomment to save video\n\n# Get input size and output size from the environment\nnactions = env.action_space.n\nninputs = env.reset().size\n# Hidden layers are arbitrarily added\n# hidden = [20, 10, 20]\nhidden = [] # ... but unnecessary with the CartPole\nnet_struct = [ninputs, *hidden, nactions]\n\n# Network setup is straightforward (defaults: `act_fn=np.tanh, init_weights=None`)\nnet = tinynet.FFNN(net_struct) # also try `RNN(net_struct)`\n\n# Get random seed for deterministic fitness (for simplicity)\nrseed = np.random.randint(1e10)\n\n# Fitness function: gameplay loop\ndef fitness(ind, render=False):\n    env.seed(rseed)  # makes fitness deterministic\n    obs = env.reset()\n    score = 0\n    done = False\n    while not done:\n      if render: env.render()\n      action = net.activate(obs).argmax()\n      obs, rew, done, info = env.step(action)\n      score += rew\n    print(f\"Score: {score}\")\n    return score\n\n# RWG does not distinguish between populations and generations\nmax_ninds = 1000\n# Neuroevolution loop\nfor nind in range(max_ninds):\n    ind = np.random.randn(net.nweights)\n    net.set_weights(ind)\n    score = fitness(ind)\n    if score == 500:\n        print(f\"Game solved in {nind} trials\")\n        break\n\n# Replay winning individual\nfitness(ind, render=True)\n\n# You may want to drop into a console here to examine the results\nimport IPython; IPython.embed()\n```",
        "description_content_type": "text/markdown",
        "docs_url": null,
        "download_url": "https://github.com/giuse/tinynet/archive/0.1.1.tar.gz",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/giuse/tinynet",
        "keywords": "neuralnetwork machinelearning",
        "license": "MIT",
        "maintainer": "",
        "maintainer_email": "",
        "name": "tinynet",
        "package_url": "https://pypi.org/project/tinynet/",
        "platform": "",
        "project_url": "https://pypi.org/project/tinynet/",
        "project_urls": {
            "Bug Reports": "https://github.com/giuse/tinynet/issues",
            "Download": "https://github.com/giuse/tinynet/archive/0.1.1.tar.gz",
            "Homepage": "https://github.com/giuse/tinynet",
            "Source": "https://github.com/giuse/tinynet/"
        },
        "release_url": "https://pypi.org/project/tinynet/0.1.1/",
        "requires_dist": null,
        "requires_python": ">=3.6, <4",
        "summary": "A tiny neural network library",
        "version": "0.1.1"
    },
    "last_serial": 5545351,
    "releases": {
        "0.0.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "e1c4c7b72633e205440132c98029086a",
                    "sha256": "ac83263f5aa1e0e1621d991487045b1f2b1a091caa4eff8c9a530dfae27c71b0"
                },
                "downloads": -1,
                "filename": "tinynet-0.0.0.tar.gz",
                "has_sig": false,
                "md5_digest": "e1c4c7b72633e205440132c98029086a",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 1627,
                "upload_time": "2019-07-02T12:04:03",
                "url": "https://files.pythonhosted.org/packages/63/c7/1f2d9b26d24b845aca499d3bea2567cf9ee5e72b47545e137b73d7a6839d/tinynet-0.0.0.tar.gz"
            }
        ],
        "0.0.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "3fa0de982a9718cfacd73b854cdf7326",
                    "sha256": "38873e4d637687718b00b57f748d067822893eb0466422f2ac6980ff5284b731"
                },
                "downloads": -1,
                "filename": "tinynet-0.0.1.tar.gz",
                "has_sig": false,
                "md5_digest": "3fa0de982a9718cfacd73b854cdf7326",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 1709,
                "upload_time": "2019-07-02T12:31:46",
                "url": "https://files.pythonhosted.org/packages/08/27/8b056ab9ca7ab27f5be252d4998b26b765c03ac3cac3b067a155be9003bf/tinynet-0.0.1.tar.gz"
            }
        ],
        "0.0.5": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "a6090cba663ed92a967d9466375f54d1",
                    "sha256": "1c99d027ba54e38ce06e6b14e000d9b8c78059a03d3556bb18a29c7535654069"
                },
                "downloads": -1,
                "filename": "tinynet-0.0.5.tar.gz",
                "has_sig": false,
                "md5_digest": "a6090cba663ed92a967d9466375f54d1",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.6, <4",
                "size": 4328,
                "upload_time": "2019-07-04T10:55:06",
                "url": "https://files.pythonhosted.org/packages/6d/d0/85a1f1c3b384a48691a8a41f545e220d4e9c8d7b9dce699c47f0bfbbaf6d/tinynet-0.0.5.tar.gz"
            }
        ],
        "0.1.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "8ee48d4ca1565913597d571c3fcfbac6",
                    "sha256": "5d3e8552e5915cb01f2ccfa537e06f53a7b6c0076217e79fca5a3bbbaf3112ab"
                },
                "downloads": -1,
                "filename": "tinynet-0.1.0.tar.gz",
                "has_sig": false,
                "md5_digest": "8ee48d4ca1565913597d571c3fcfbac6",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.6, <4",
                "size": 4699,
                "upload_time": "2019-07-16T21:09:44",
                "url": "https://files.pythonhosted.org/packages/9b/42/b3f1d6ec6e2e6f1b29acf9267a1252a882cae844f47efb800968d94e51f2/tinynet-0.1.0.tar.gz"
            }
        ],
        "0.1.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "505db116bb1efc71ba67b5c931d2df4b",
                    "sha256": "67dc9789bb11c48731dfbb8894e265c18058a1a614dd45abc1f2ce27a307c1d4"
                },
                "downloads": -1,
                "filename": "tinynet-0.1.1.tar.gz",
                "has_sig": false,
                "md5_digest": "505db116bb1efc71ba67b5c931d2df4b",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">=3.6, <4",
                "size": 5045,
                "upload_time": "2019-07-17T11:23:40",
                "url": "https://files.pythonhosted.org/packages/33/0b/b0b2c015b5411161a6d76ab76d046eed9cd4c01c5c7340cda60f770e28b8/tinynet-0.1.1.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "505db116bb1efc71ba67b5c931d2df4b",
                "sha256": "67dc9789bb11c48731dfbb8894e265c18058a1a614dd45abc1f2ce27a307c1d4"
            },
            "downloads": -1,
            "filename": "tinynet-0.1.1.tar.gz",
            "has_sig": false,
            "md5_digest": "505db116bb1efc71ba67b5c931d2df4b",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": ">=3.6, <4",
            "size": 5045,
            "upload_time": "2019-07-17T11:23:40",
            "url": "https://files.pythonhosted.org/packages/33/0b/b0b2c015b5411161a6d76ab76d046eed9cd4c01c5c7340cda60f770e28b8/tinynet-0.1.1.tar.gz"
        }
    ]
}