A short description of the project:
- This is the group repo for 02476 Machine Learning Operation Jan. 2022.
- Course Home page
- FrameWork Transformers
- To get setup with both code and data by simply running:
git clone https://github.com/AbdulstarKousa/Project-MLOps-2022.gitdvc pull
- To pull the docker image from the gcp server:
docker pull gcr.io/team-x-338109/main:latest
- To run the docker image interactively:
docker run -it gcr.io/team-x-338109/main:latest
A trained deployable model is provided in a docker image
- To pull the deployable image from gcp run the following command:
docker pull gcr.io/team-x-338109/deploy:latest
- To deploy the model locally run the docker image using the following command:
docker run -d -p 8080:8080 --name=local_deploy gcr.io/team-x-338109/deploy:latest
- A text file with a review can be provided for inference using:
curl localhost:8080/predictions/distilbert -T review.txt
In this project we make use of the Hugging Face Transformer framework to create a binary sentiment classifier for review’s given to highly polarized movies on IMDb. The dataset is available on the Hugging Face Github page.
The model used for performing the classification is a Bidirectional Encoder Representations from Transformers also known as BERT. Transfer learning will be applied by training a pre-trained model on 25.000 positive/negative labeled movie reviews, and then evaluated on another 25.000 reviews. Different configurations of BERT will be trained and evaluated.
The whole pipeline is also available as a Docker-Image:
├── LICENSE
├── Makefile <- Makefile with commands like `make data` or `make train`
├── README.md <- The top-level README for developers using this project.
├── data
│ ├── external <- Data from third party sources.
│ ├── interim <- Intermediate data that has been transformed.
│ ├── processed <- The final, canonical data sets for modeling.
│ └── raw <- The original, immutable data dump.
│
├── docs <- A default Sphinx project; see sphinx-doc.org for details
│
├── models <- Trained and serialized models, model predictions, or model summaries
│
├── notebooks <- Jupyter notebooks. Naming convention is a number (for ordering),
│ the creator's initials, and a short `-` delimited description, e.g.
│ `1.0-jqp-initial-data-exploration`.
│
├── references <- Data dictionaries, manuals, and all other explanatory materials.
│
├── reports <- Generated analysis as HTML, PDF, LaTeX, etc.
│ └── figures <- Generated graphics and figures to be used in reporting
│
├── requirements.txt <- The requirements file for reproducing the analysis environment, e.g.
│ generated with `pip freeze > requirements.txt`
│
├── setup.py <- makes project pip installable (pip install -e .) so src can be imported
├── src <- Source code for use in this project.
│ ├── __init__.py <- Makes src a Python module
│ │
│ ├── data <- Scripts to download or generate data
│ │ └── make_dataset.py
│ │
│ ├── features <- Scripts to turn raw data into features for modeling
│ │ └── build_features.py
│ │
│ ├── models <- Scripts to train models and then use trained models to make
│ │ │ predictions
│ │ ├── predict_model.py
│ │ └── train_model.py
│ │
│ └── visualization <- Scripts to create exploratory and results oriented visualizations
│ └── visualize.py
│
└── tox.ini <- tox file with settings for running tox; see tox.readthedocs.io
This project is licensed under the MIT License. You are free to use the code and resources for educational or personal purposes.
I welcome any feedback, suggestions, or questions you may have about the projects or the repository. Feel free to reach out to me via email at abdul.kousa@gmail.com
Project based on the cookiecutter data science project template. #cookiecutterdatascience