{
    "info": {
        "author": "Ivan Itzcovich",
        "author_email": "i.itzcovich@gmail.com",
        "bugtrack_url": null,
        "classifiers": [
            "License :: OSI Approved :: MIT License",
            "Programming Language :: Python :: 3.5",
            "Programming Language :: Python :: 3.6",
            "Programming Language :: Python :: 3.7",
            "Topic :: Internet :: WWW/HTTP :: HTTP Servers",
            "Topic :: Internet :: WWW/HTTP :: WSGI",
            "Topic :: Internet :: WWW/HTTP :: WSGI :: Application",
            "Topic :: Internet :: WWW/HTTP :: WSGI :: Middleware",
            "Topic :: Internet :: WWW/HTTP :: WSGI :: Server",
            "Topic :: Software Development :: Libraries :: Application Frameworks"
        ],
        "description": "# TFServe\n\n[![Downloads](https://pepy.tech/badge/tfserve)](https://pepy.tech/project/tfserve)  [![PyPI version](https://badge.fury.io/py/tfserve.svg)](https://badge.fury.io/py/tfserve)\n\nTFServe is a framework designed to serve tensorflow models in a simple and easy way as an HTTP API server. It's built on top of [Werkzeug](http://werkzeug.pocoo.org/).\n\n## How to install\n\n```bash\n$ pip install tfserve\n```\n\nAfter installing `tfserve`, install either `tensorflow` of `tensorflow-gpu` (the latter if you have GPU available).\n\n```bash\n$ pip install tensorflow\n```\nor\n```bash\n$ pip install tensorflow-gpu\n```\n\n## How to use\n\n### Python API\n\nYou will need 5 parts:\n\n1. **Model**: it can be a `.pb` file or a model directory containing ckpt files.\n2. **Input tensor names**: name of the input tensors of the graph.\n3. **Output tensor names**: name of the output tensors of the graph.\n4. **`encode`**: python function that receives the request body data and outputs a `dict` mapping input tensor names to input numpy values.\n5. **`decode`**: python function that receives a `dict` mapping output tensor names to output numpy values and returns the HTTP response.\n\nFollow the example to learn how to combine these parts...\n\n#### Example\n\nDeploy image classification service that receives a binary jpg image and returns the class of the object found in the image alongside it's probability.\n\n```python\n\n# 1. Model: trained mobilenet on ImageNet that can be downloaded from\n#           https://storage.googleapis.com/mobilenet_v2/checkpoints/mobilenet_v2_1.4_224.tgz\nMODEL_PATH = \"mobilenet_v2_1.4_224/mobilenet_v2_1.4_224_frozen.pb\"\n\n# 2. Input tensor names:\nINPUT_TENSORS = [\"import/input:0\"]\n\n# 3. Output tensor names:\nOUTPUT_TENSORS = [\"import/MobilenetV2/Predictions/Softmax:0\"]\n\n# 4. encode function: Receives raw jpg image as request data. Returns dict\n#                     mappint import/input:0 to numpy value.\ndef encode(request_data):\n    with tempfile.NamedTemporaryFile(mode=\"wb\", suffix=\".jpg\") as f:\n        f.write(request_data)\n        # Model receives 224x224 normalized RGB image.\n        img = Image.open(f.name).resize((224, 224)) \n        img = np.asarray(img) / 255.\n\n    return {INPUT_TENSORS[0]: img}\n\n# 5. decode function: Receives `dict` mapping import/MobilenetV2/Predictions/Softmax:0 to\n#                     numpy value and builds dict with for json response.\ndef decode(outputs):\n    p = outputs[OUTPUT_TENSORS[0]] # 1001 vector with probabilities for each class.\n    index = np.argmax(p)\n    return {\"class\": index_to_class_map[index-1], \"prob\": float(p[index])}\n```\n\nThat's it! Now create TFServeApp object and run it!\n\n```python\napp = TFServeApp(MODEL_PATH, INPUT_TENSORS, OUTPUT_TENSORS, encode, decode)\napp.run('127.0.0.1', 5000)  # Host and port where the server will be running\n```\n\nSee `client.py` for full example.\n\n#### How to consume server\n\n![img](imgs/screen.gif)\n\n> The server supports only `POST` method to `/` with the input information as part of the request body.\n\nThe input will be proccessed in the encode function to produce the feed_dict object that will be passed to the graph. The graph output will be processed in the decode function and the server will return whatever the decode function returns.\n\n### CLI\n\n`tfserve` also provides a CLI program with built-in encode/decode handlers:\n\n```bash\ntfserve -m PATH [-i INPUTS] [-o OUTPUTS] [-h HANDLER] [-b] [-H HOST] [-p PORT]\n\n  -m PATH, --model PATH\n                        path to pb file or directory containing checkpoint\n  -i INPUTS, --inputs INPUTS\n                        a comma separated list of input tensors\n  -o OUTPUTS, --outputs OUTPUTS\n                        a comma separated list of output tensors\n  -h HANDLER, --handler HANDLER\n                        encode/decode handler (deault is 'json')\n  -b, --batch           process multiple inputs (default is to process\n                        one input per request)\n  -H HOST, --host HOST  host interface to bind to (0.0.0.0)\n  -p PORT, --port PORT  port to listen on (5000)\n```\n\n#### Example\n\n```bash\n$ tfserve -m models/graph.pb -i x:0 -o out:0 -h json -H localhost\n```\n\nRun `tfserve` with `models/graph.pb` model that takes as input the tensor with name `x:0` (dimesion: [?,5]) and outputs a tensor named `out:0`. The server will run on http://localhost:5000/ and will receive `POST` requests to `/`.\n\nBy using the **json handler**, you can provide the input data as a `json` object in the request body:\n\n```json\n{\n  \"x:0\": [1,1,3,4,5]\n}\n```\n\nYou will receive a `json` output object as:\n\n```json\n{\n  \"out:0\": 0.48\n}\n```\n\n#### More information about CLI\n\nRun:\n\n```bash\n$ tfserve --help\n```\n\n## Help\n\n* **What if I don't know the tensor names?**\n\n> You can use `tfserve.helper.estimate_io_tensors(model_path)` function to get a list of possible input/output tensor names. Also, you can use the CLI by running: `tfserve -m [model_path] --help-model`\n\n* **What if I want to run multiple inferences at the same time?**\n\n> You can use `batch=True` when building tfserve.TFServeApp. You will then need to handle the batch dimension yourself in the `encode` and `decode` function.\n> Also, if using the CLI, just add the `--batch` flag.\n\n\n## Limitation\n\n> It only works with one-to-one models. That is, models that need to run the graph only once to get the inference.\n> Other architectures of inference will be supported soon. Help is appreciated!\n\n\n",
        "description_content_type": "text/markdown",
        "docs_url": null,
        "download_url": "",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "http://github.com/iitzco/tfserve",
        "keywords": "tensorflow deep-learning serving",
        "license": "MIT",
        "maintainer": "",
        "maintainer_email": "",
        "name": "tfserve",
        "package_url": "https://pypi.org/project/tfserve/",
        "platform": "any",
        "project_url": "https://pypi.org/project/tfserve/",
        "project_urls": {
            "Homepage": "http://github.com/iitzco/tfserve"
        },
        "release_url": "https://pypi.org/project/tfserve/0.3/",
        "requires_dist": [
            "numpy",
            "apistar (==0.5.41)",
            "nose",
            "Werkzeug"
        ],
        "requires_python": ">3.5",
        "summary": "Serve TF models simple and easy as an HTTP API server.",
        "version": "0.3"
    },
    "last_serial": 4428167,
    "releases": {
        "0.1.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "b729f66472df7e03d573aad106ec7c13",
                    "sha256": "b803a13a45f4487ec9a61973aa998f452e4f595597f02b20748471aeef327103"
                },
                "downloads": -1,
                "filename": "tfserve-0.1.0-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "b729f66472df7e03d573aad106ec7c13",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": ">3.5",
                "size": 7514,
                "upload_time": "2018-09-05T16:34:48",
                "url": "https://files.pythonhosted.org/packages/ba/38/ecae7921edb4e72a7dfe12b9ab570d4faf100e3c4f96fb2f5f6406c9632f/tfserve-0.1.0-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "b700daf4eb2d6cf52e50ade5a504016c",
                    "sha256": "e7e8f207bb19a4bd6eeb4a5e04cc7fa2c32f1d69084c777dc2ff75b9a6c01c98"
                },
                "downloads": -1,
                "filename": "tfserve-0.1.0.tar.gz",
                "has_sig": false,
                "md5_digest": "b700daf4eb2d6cf52e50ade5a504016c",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">3.5",
                "size": 11561,
                "upload_time": "2018-09-05T16:34:50",
                "url": "https://files.pythonhosted.org/packages/17/e7/242bc4ba9e3e81ee22e22022ee173a91f384d914f42f6bdd55d5d84ffcef/tfserve-0.1.0.tar.gz"
            }
        ],
        "0.1.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "1dac5fdda860b9f8d5d7e477b8f111f2",
                    "sha256": "bcefd145c6d6197979b3b1c6d883ee96be1498a9b41e8ef12916603acf27df14"
                },
                "downloads": -1,
                "filename": "tfserve-0.1.1-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "1dac5fdda860b9f8d5d7e477b8f111f2",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": ">3.5",
                "size": 7519,
                "upload_time": "2018-10-10T21:43:07",
                "url": "https://files.pythonhosted.org/packages/43/43/5a48cddbdbca7fcbdcc800872b639bca527c062de150b9a6fd149c759b26/tfserve-0.1.1-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "fef8a76a3858bc6aed6f52724e6d536c",
                    "sha256": "842439d0d9f32bf1c937855f0b9587eadda740424d184499fec8b151dedeb219"
                },
                "downloads": -1,
                "filename": "tfserve-0.1.1.tar.gz",
                "has_sig": false,
                "md5_digest": "fef8a76a3858bc6aed6f52724e6d536c",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">3.5",
                "size": 11564,
                "upload_time": "2018-10-10T21:43:10",
                "url": "https://files.pythonhosted.org/packages/e0/dc/a0f61e63436fe75b83ae0f1483f58ca23c9712109790435e89fd1aee718c/tfserve-0.1.1.tar.gz"
            }
        ],
        "0.2": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "db0e4affc4af5dc55fb4cccf6f10e566",
                    "sha256": "07b9f8949a5ebc41435844c8492ee3a4be7bd14435b7a815f338d7d5289dfb19"
                },
                "downloads": -1,
                "filename": "tfserve-0.2-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "db0e4affc4af5dc55fb4cccf6f10e566",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": ">3.5",
                "size": 8788,
                "upload_time": "2018-10-16T14:12:40",
                "url": "https://files.pythonhosted.org/packages/46/70/d7411274efa9b1665b0c2129e0f541afd947ef8438226350bcea87d53683/tfserve-0.2-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "da96bad9259dc58ed9a3519a3c810aec",
                    "sha256": "57465ab7202b35934ca085a154c6b7433025b7bb07efbe3940e660213566390a"
                },
                "downloads": -1,
                "filename": "tfserve-0.2.tar.gz",
                "has_sig": false,
                "md5_digest": "da96bad9259dc58ed9a3519a3c810aec",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">3.5",
                "size": 12767,
                "upload_time": "2018-10-16T14:12:42",
                "url": "https://files.pythonhosted.org/packages/4d/a8/f641c4ef8006aec1b14122d28e26d245a1df2a5f9d457cde06d36fe88500/tfserve-0.2.tar.gz"
            }
        ],
        "0.2.1": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "007dd19bd2f430a04698689983b65908",
                    "sha256": "0297529a52ec3e062ad90c7466b981cb79e4bb2a40d8f846bb6dcbbdd0195bec"
                },
                "downloads": -1,
                "filename": "tfserve-0.2.1-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "007dd19bd2f430a04698689983b65908",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": ">3.5",
                "size": 8766,
                "upload_time": "2018-10-16T14:21:34",
                "url": "https://files.pythonhosted.org/packages/27/10/068d8a85e86e6b10529174626a35a0780ce39a2f2d94cdf130afb7f8995c/tfserve-0.2.1-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "aebd7e96924c7a1a120c80e3c0c55bd8",
                    "sha256": "e7e862f6e42596df5710404fd7d6733787afa601862616d00b562cdc042376d0"
                },
                "downloads": -1,
                "filename": "tfserve-0.2.1.tar.gz",
                "has_sig": false,
                "md5_digest": "aebd7e96924c7a1a120c80e3c0c55bd8",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">3.5",
                "size": 12844,
                "upload_time": "2018-10-16T14:21:35",
                "url": "https://files.pythonhosted.org/packages/9e/5a/6e7fc35935f86ed47ebab1586fb9b40076ca9cb97571ee1ac8af0c5a7ebf/tfserve-0.2.1.tar.gz"
            }
        ],
        "0.3": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "309ad43bdfafc86f46eafe277777ecc4",
                    "sha256": "201a091a80722aac5bfa7d7842e794c260f9aee84bf127cffadbc3b9bd845f31"
                },
                "downloads": -1,
                "filename": "tfserve-0.3-py3-none-any.whl",
                "has_sig": false,
                "md5_digest": "309ad43bdfafc86f46eafe277777ecc4",
                "packagetype": "bdist_wheel",
                "python_version": "py3",
                "requires_python": ">3.5",
                "size": 13333,
                "upload_time": "2018-10-29T16:06:04",
                "url": "https://files.pythonhosted.org/packages/47/91/f6b59699d0d18424df8926197544e1143f791e8b19c5883f138c8eaa604a/tfserve-0.3-py3-none-any.whl"
            },
            {
                "comment_text": "",
                "digests": {
                    "md5": "52139a92094105929eb40b7b26e2e541",
                    "sha256": "11dbda4303a3e8955332d4b60aa81c8a8f77428abee2e4170838d1e4f687fdf8"
                },
                "downloads": -1,
                "filename": "tfserve-0.3.tar.gz",
                "has_sig": false,
                "md5_digest": "52139a92094105929eb40b7b26e2e541",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": ">3.5",
                "size": 16922,
                "upload_time": "2018-10-29T16:06:08",
                "url": "https://files.pythonhosted.org/packages/cd/ec/630490066f3a56d1c5be7a963e3bed829f4bd8e0df507bacb729eef6655a/tfserve-0.3.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "309ad43bdfafc86f46eafe277777ecc4",
                "sha256": "201a091a80722aac5bfa7d7842e794c260f9aee84bf127cffadbc3b9bd845f31"
            },
            "downloads": -1,
            "filename": "tfserve-0.3-py3-none-any.whl",
            "has_sig": false,
            "md5_digest": "309ad43bdfafc86f46eafe277777ecc4",
            "packagetype": "bdist_wheel",
            "python_version": "py3",
            "requires_python": ">3.5",
            "size": 13333,
            "upload_time": "2018-10-29T16:06:04",
            "url": "https://files.pythonhosted.org/packages/47/91/f6b59699d0d18424df8926197544e1143f791e8b19c5883f138c8eaa604a/tfserve-0.3-py3-none-any.whl"
        },
        {
            "comment_text": "",
            "digests": {
                "md5": "52139a92094105929eb40b7b26e2e541",
                "sha256": "11dbda4303a3e8955332d4b60aa81c8a8f77428abee2e4170838d1e4f687fdf8"
            },
            "downloads": -1,
            "filename": "tfserve-0.3.tar.gz",
            "has_sig": false,
            "md5_digest": "52139a92094105929eb40b7b26e2e541",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": ">3.5",
            "size": 16922,
            "upload_time": "2018-10-29T16:06:08",
            "url": "https://files.pythonhosted.org/packages/cd/ec/630490066f3a56d1c5be7a963e3bed829f4bd8e0df507bacb729eef6655a/tfserve-0.3.tar.gz"
        }
    ]
}