{ "info": { "author": "Alec Radford, Madison May", "author_email": "Alec Radford ,\n Madison May ", "bugtrack_url": null, "classifiers": [], "description": "Passage\n=======\n\nA little library for text analysis with RNNs.\n\nWarning: very alpha, work in progress.\n\nInstall\n-------\n\nvia Github (version under active development)\n\n::\n\n git clone http://github.com/IndicoDataSolutions/passage.git\n python setup.py develop\n\nor via pip\n\n::\n\n sudo pip install passage\n\nExample\n-------\n\nUsing Passage to do binary classification of text, this example:\n\n- Tokenizes some training text, converting it to a format Passage can\n use.\n- Defines the model's structure as a list of layers.\n- Creates the model with that structure and a cost to be optimized.\n- Trains the model for one iteration over the training text.\n- Uses the model and tokenizer to predict on new text.\n- Saves and loads the model.\n\n::\n\n from passage.preprocessing import Tokenizer\n from passage.layers import Embedding, GatedRecurrent, Dense\n from passage.models import RNN\n from passage.utils import save, load\n\n tokenizer = Tokenizer()\n train_tokens = tokenizer.fit_transform(train_text)\n\n layers = [\n Embedding(size=128, n_features=tokenizer.n_features),\n GatedRecurrent(size=128),\n Dense(size=1, activation='sigmoid')\n ]\n\n model = RNN(layers=layers, cost='BinaryCrossEntropy')\n model.fit(train_tokens, train_labels)\n\n model.predict(tokenizer.transform(test_text))\n save(model, 'save_test.pkl')\n model = load('save_test.pkl')\n\nWhere:\n\n- train\\_text is a list of strings ['hello world', 'foo bar']\n- train\\_labels is a list of labels [0, 1]\n- test\\_text is another list of strings\n\nDatasets\n--------\n\nWithout sizeable datasets RNNs have difficulty achieving results better\nthan traditional sparse linear models. Below are a few datasets that are\nappropriately sized, useful for experimentation. Hopefully this list\nwill grow over time, please feel free to propose new datasets for\ninclusion through either an issue or a pull request.\n\n****Note****: **None of these datasets were created by indico, not\nshould their inclusion here indicate any kind of endorsement**\n\nBlogger Dataset: http://www.cs.biu.ac.il/~koppel/blogs/blogs.zip (Age\nand gender data)", "description_content_type": null, "docs_url": null, "download_url": "UNKNOWN", "downloads": { "last_day": -1, "last_month": -1, "last_week": -1 }, "home_page": "https://github.com/IndicoDataSolutions/Passage", "keywords": null, "license": "MIT License (See LICENSE)", "maintainer": null, "maintainer_email": null, "name": "passage", "package_url": "https://pypi.org/project/passage/", "platform": "UNKNOWN", "project_url": "https://pypi.org/project/passage/", "project_urls": { "Download": "UNKNOWN", "Homepage": "https://github.com/IndicoDataSolutions/Passage" }, "release_url": "https://pypi.org/project/passage/0.2.4/", "requires_dist": null, "requires_python": null, "summary": "A little library for text analysis with RNNs.", "version": "0.2.4" }, "last_serial": 1436376, "releases": { "0.2.1": [ { "comment_text": "", "digests": { "md5": "504aab72a0c0263b6463d4c8932affac", "sha256": "45af592c0ac9dc124a97d4fdd5d19871a9f82f384d444194d29b0f680a92bda2" }, "downloads": -1, "filename": "passage-0.2.1.tar.gz", "has_sig": false, "md5_digest": "504aab72a0c0263b6463d4c8932affac", "packagetype": "sdist", "python_version": "source", "requires_python": null, "size": 8184, "upload_time": "2015-02-16T23:52:54", "url": "https://files.pythonhosted.org/packages/b3/79/8506d4d5e3afaeb5a1164794bc87d6288c2492100db7935c0f2459396b63/passage-0.2.1.tar.gz" } ], "0.2.2": [ { "comment_text": "", "digests": { "md5": "2cb53b0335c11527d3cf45984bf608ee", "sha256": "f941a0a9146089b73dc920ad0c01d175eca31e08298e9120ecaf8d6b5bc7d14d" }, "downloads": -1, "filename": "passage-0.2.2.tar.gz", "has_sig": false, "md5_digest": "2cb53b0335c11527d3cf45984bf608ee", "packagetype": "sdist", "python_version": "source", "requires_python": null, "size": 9434, "upload_time": "2015-02-24T16:15:46", "url": "https://files.pythonhosted.org/packages/cb/56/df27cb1b7f0f416dc103e5571d9a4e2e817c2d8ff243e035486ccff0e56d/passage-0.2.2.tar.gz" } ], "0.2.3": [ { "comment_text": "", "digests": { "md5": "9b0e6542ea5b6e142fc155bea6b4e96b", "sha256": "200fbb2492c3a9a14d11cca3a9c8563ca84e7b564943b0a8293e5fbf0d10bc63" }, "downloads": -1, "filename": "passage-0.2.3.tar.gz", "has_sig": false, "md5_digest": "9b0e6542ea5b6e142fc155bea6b4e96b", "packagetype": "sdist", "python_version": "source", "requires_python": null, "size": 8519, "upload_time": "2015-02-24T16:21:32", "url": "https://files.pythonhosted.org/packages/04/fd/b273054b8245e018a037b8393316bf84de117d4c9a29f062cd5ccecc10a7/passage-0.2.3.tar.gz" } ], "0.2.4": [ { "comment_text": "", "digests": { "md5": "3496f6e1226d11c39ee8af23dfc3d224", "sha256": "45f9113b3e5b5fe3f1452888ceacde99009a07b2e564f74fe2f0117987d723a3" }, "downloads": -1, "filename": "passage-0.2.4.tar.gz", "has_sig": false, "md5_digest": "3496f6e1226d11c39ee8af23dfc3d224", "packagetype": "sdist", "python_version": "source", "requires_python": null, "size": 9708, "upload_time": "2015-02-24T17:15:55", "url": "https://files.pythonhosted.org/packages/6a/5b/65dbad95c7195954f20a9f88d406274bcecdf9d373f4387097a1a7acb69d/passage-0.2.4.tar.gz" } ] }, "urls": [ { "comment_text": "", "digests": { "md5": "3496f6e1226d11c39ee8af23dfc3d224", "sha256": "45f9113b3e5b5fe3f1452888ceacde99009a07b2e564f74fe2f0117987d723a3" }, "downloads": -1, "filename": "passage-0.2.4.tar.gz", "has_sig": false, "md5_digest": "3496f6e1226d11c39ee8af23dfc3d224", "packagetype": "sdist", "python_version": "source", "requires_python": null, "size": 9708, "upload_time": "2015-02-24T17:15:55", "url": "https://files.pythonhosted.org/packages/6a/5b/65dbad95c7195954f20a9f88d406274bcecdf9d373f4387097a1a7acb69d/passage-0.2.4.tar.gz" } ] }