{
    "info": {
        "author": "Jonathan Raiman",
        "author_email": "jonathanraiman@gmail.com",
        "bugtrack_url": null,
        "classifiers": [
            "Intended Audience :: Science/Research",
            "Operating System :: OS Independent",
            "Programming Language :: Python :: 3.3"
        ],
        "description": "# Glove\n\nCython general implementation of the Glove multi-threaded training.\n\nGloVe is an unsupervised learning algorithm for generating vector representations for words.\nTraining is done using a co-occcurence matrix from a corpus. The resulting representations contain structure useful for many other tasks.\n\nThe paper describing the model is [here](http://nlp.stanford.edu/projects/glove/glove.pdf).\n\nThe original implementation for this Machine Learning model can be [found here](http://nlp.stanford.edu/projects/glove/).\n\n@author Jonathan Raiman\n\n## Example\n\nTo use this package you need a sparse co-occurence matrix.\nThis matrix is represented by nested dictionaries that use ints as keys\nwith a 0-index.\n\nFor instance below we have a corpus of 3 indices. Below 0 co-occurs with 2, 3.5 times:\n\n```python\nimport glove\n\ncooccur = {\n\t0: {\n\t\t0: 1.0,\n\t\t2: 3.5\n\t},\n\t1: {\n\t\t2: 0.5\n\t},\n\t2: {\n\t\t0: 3.5,\n\t\t1: 0.5,\n\t\t2: 1.2\n\t}\n}\n\nmodel = glove.Glove(cooccur, vocab_size=3, d=50, alpha=0.75, x_max=100.0)\n\nfor epoch in range(25):\n    err = model.train(batch_size=200, workers=9, batch_size=50)\n    print(\"epoch %d, error %.3f\" % (epoch, err), flush=True)\n```\n\nThe trained embeddings are now present under `model.W`.\n\n## Usage\n\nThe model is controlled by setting several hyperpameters.\n\n### Glove.__init__()\n\n* `cooccurence` dict<int, dict<int, float>> : the co-occurence matrix\n* `alpha` float : (default 0.75) hyperparameter for controlling the exponent for normalized co-occurence counts.\n* `x_max` float : (default 100.0) hyperparameter for controlling smoothing for common items in co-occurence matrix.\n* `d` int : (default 50) how many embedding dimensions for learnt vectors\n* `seed` int : (default 1234) the random seed\n\n### Glove.train\n\n* `step_size` float : the learning rate for the model\n* `workers` int : number of worker threads used for training\n* `batch_size` int : how many examples should each thread receive (controls the size of the job queue)",
        "description_content_type": null,
        "docs_url": null,
        "download_url": "https://github.com/JonathanRaiman/glove",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "https://github.com/JonathanRaiman/glove",
        "keywords": "NLP,Machine Learning",
        "license": "MIT",
        "maintainer": "",
        "maintainer_email": "",
        "name": "glove",
        "package_url": "https://pypi.org/project/glove/",
        "platform": "any",
        "project_url": "https://pypi.org/project/glove/",
        "project_urls": {
            "Download": "https://github.com/JonathanRaiman/glove",
            "Homepage": "https://github.com/JonathanRaiman/glove"
        },
        "release_url": "https://pypi.org/project/glove/1.0.2/",
        "requires_dist": null,
        "requires_python": "",
        "summary": "Python package for computing embeddings from co-occurence matrices",
        "version": "1.0.2"
    },
    "last_serial": 3287245,
    "releases": {
        "1.0.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "0fcbf26d5585a6081df18508d1f147d6",
                    "sha256": "c639bd590bef579fa9dedaf03e78f54e0c9f285436c925a3b5c26466c44f1fe6"
                },
                "downloads": -1,
                "filename": "glove-1.0.0.tar.gz",
                "has_sig": false,
                "md5_digest": "0fcbf26d5585a6081df18508d1f147d6",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 3598,
                "upload_time": "2015-09-21T15:34:22",
                "url": "https://files.pythonhosted.org/packages/2a/d5/0c0a5e4ec274ab6e18e053a7af757c57d91a3eaedc27a6a2a3c0578bcdff/glove-1.0.0.tar.gz"
            }
        ],
        "1.0.2": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "7890930e2d401b63f0079ee812979ea2",
                    "sha256": "b2e00cdcc3fa77a72f4e6ab89f73236da34feb0e38908b5aea8110cdb3b747c6"
                },
                "downloads": -1,
                "filename": "glove-1.0.2.tar.gz",
                "has_sig": false,
                "md5_digest": "7890930e2d401b63f0079ee812979ea2",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 44867,
                "upload_time": "2017-10-29T04:15:21",
                "url": "https://files.pythonhosted.org/packages/8a/c9/17c400d0c29746162bd47fc719bf3212b2b031949d41d712e9bdef11ae03/glove-1.0.2.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "7890930e2d401b63f0079ee812979ea2",
                "sha256": "b2e00cdcc3fa77a72f4e6ab89f73236da34feb0e38908b5aea8110cdb3b747c6"
            },
            "downloads": -1,
            "filename": "glove-1.0.2.tar.gz",
            "has_sig": false,
            "md5_digest": "7890930e2d401b63f0079ee812979ea2",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 44867,
            "upload_time": "2017-10-29T04:15:21",
            "url": "https://files.pythonhosted.org/packages/8a/c9/17c400d0c29746162bd47fc719bf3212b2b031949d41d712e9bdef11ae03/glove-1.0.2.tar.gz"
        }
    ]
}