https://github.com/sbcd90/machine-learning-rest-server
This project implements a common rest server which can serve tensorflow-serving & xgboost models.
https://github.com/sbcd90/machine-learning-rest-server
machine-learning proxygen rest serving tensorflow xgboost
Last synced: 8 months ago
JSON representation
This project implements a common rest server which can serve tensorflow-serving & xgboost models.
- Host: GitHub
- URL: https://github.com/sbcd90/machine-learning-rest-server
- Owner: sbcd90
- Created: 2018-07-15T22:25:02.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2018-07-15T22:27:47.000Z (over 7 years ago)
- Last Synced: 2025-04-26T20:33:55.788Z (10 months ago)
- Topics: machine-learning, proxygen, rest, serving, tensorflow, xgboost
- Language: C++
- Size: 75.2 KB
- Stars: 5
- Watchers: 1
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
ml_model_rest_server
====================
This project implements a common rest server which can serve tensorflow-serving & xgboost models.
Architecture
============

Installation
============
### Dependencies
- [proxygen](https://github.com/facebook/proxygen)
- [tensorflow](https://github.com/tensorflow/tensorflow)
- [tensorflow-serving](https://github.com/tensorflow/serving)
- [xgboost](https://github.com/dmlc/xgboost)
### Steps
- ./configure
- cd debug
- make
Example
=======
- To serve tf-serving model, start `tf-serving` first.
```
POST http://localhost:11000
{"input_type": "ints", "inputs": [1, 2, 4, 5], "model_name": "tensorflow-feature-sum-model", "model_signature_name": "tensorflow-feature-sum-model"}
```
- Output
```
{"y_output":"12"}
```
- To serve xgboost models, put the `*.cpb` files in the `debug/src/frontends` directory.
```
POST http://localhost:11000
{"input_type": "floats", "inputs": [1.0, 2.0, 4.0, 5.0], "model_name": "test.cpb", "model_signature_name": "xgboost"}
```
- Output
```
{"y_output":"0.0001"}
```
Sample models
=============
A sample `xgboost test.cpb` model is present in `resources` directory.
Next steps
==========
- docker integration