/tf_serving

Primary LanguageJupyter Notebook

tf_Serving

Prerequisites

  • Python 3.x
  • Docker (for TensorFlow Serving)
  • Operating system: Ubuntu 20.04

In Hospital Mortality Prediction

  • Description: This dataset is designed for predicting in-hospital mortality and may include patient records, clinical measurements, and outcomes.

  • Data Format: [Specify the format of the dataset, e.g., CSV, JSON.]

  • Data Fields: [List and describe the key fields or columns in the dataset. You can find this information on the Kaggle dataset page.]

  • Data Usage: This dataset can be used for predicting in-hospital mortality, healthcare analytics, and machine learning research.

  • License: Please refer to the Kaggle dataset page for information regarding the dataset's license and terms of use.

  • Download Link: You can download the dataset from the Kaggle dataset page.

Usage Guidelines

[Include any specific guidelines or instructions for users who want to use the dataset, such as data preprocessing steps or ethical considerations.]

Usage

  1. setup the project:

    virtualenv venv
source venv/bin/activate

  2. run the code@

    In-hospital mortality prediction#

    Logistic Regression
     ```bash
 python in_hospital_mortality/lr/main.py --l2 --C 0.001 --output_dir in_hospital_mortality/lr/  
 ```
    Ann model
     ```bash
 python ptbdb/baseline.py
 ```
    Logistic Regression
     ```bash
 python xgboost/main.py  --output_dir in_hospital_mortality/xgboost/
 ```

    Phenotype classification

    LSTM

    Train

    python phenotyping/rnn/main.py --network models/lstm.py --dim 256 --timestep 1.0 --depth 1 --dropout 0.3 --mode train --batch_size 8 --output_dir phenotyping/
    XGBoost

    Train & Test

    python phenotyping/xgboost/main.py  --output_dir phenotyping/xgboost/
    LightGBM

    Train & Test

    python phenotyping/lightgbm/main.py  --output_dir phenotyping/lightgbm/
    XGBoost FBFE

    Train & Test

    python phenotyping/xgboost_FBFE/main.py --timestep 1.0  --output_dir phenotyping/xgboost_FBFE/

  3. Deploy into tf_serving:

    Download the TensorFlow Serving Docker image

    docker pull tensorflow/serving

    Location of demo models

    model="$(pwd)/export_tf/moritality"

    Start TensorFlow Serving container and open the REST API port

    docker run -t --rm -p 8501:8501 \
    -v "$model:/models" \
    -e MODEL_NAME=ann_baseline \
    tensorflow/serving &

  4. Query the model using the predict API

    curl -d '{"instances": [[1.0, 2.0, 5.0]]}' \
    -X POST http://localhost:8501/v1/models/cnn_baseline:predict

    Returns => { "predictions": [2.5, 3.0, 4.5] }