Hey there! Get ready to discover how they can help us better understand language! I'll be exploring two super cool techniques—LSTM pooling and weighted pooling—that will give you a deeper understanding of these models. These methods are tested in sentiment analysis tasks.
.
├── config
│ ├── main.yaml # Main configuration file
│ ├── model # Configurations for training model
│ │ ├── model1.yaml # First variation of parameters to train model
│ │ └── model2.yaml # Second variation of parameters to train model
│ └── process # Configurations for processing data
│ ├── process1.yaml # First variation of parameters to process data
│ └── process2.yaml # Second variation of parameters to process data
├── docs # documentation for your project
├── dvc.yaml # DVC pipeline
├── .flake8 # configuration for flake8 - a Python formatter tool
├── .gitignore # ignore files that cannot commit to Git
├── Makefile # store useful commands to set up the environment
├── pyproject.toml # dependencies for poetry
├── README.md # describe your project
├── src # store source code
│ ├── __init__.py # make src a Python module
│ ├── data_batcher.py # process data before training model
│ └── data_loader.py # batch the dataset
│ └── evaluate.py # evaluating during training
│ └── inference.py # inference script
│ └── main.py # trainer class
│ └── model.py # model architecure
│ └── pretrainedModel.py # download/load pretrained model
│ └── train_utils.py # train and evluate model
│ └── train.py # parse paramters to train
│ └── train.sh # parse paramters to train
│ └── utils.py # utils to train
│ └── visualize.py # visualize weights for each epoch
| Feature | Description | |
|---|---|---|
| ⚙️ | Architecture | The project leverages a Python 3.10 environment, utilizing the huggingface package for model training |
| 🔩 | Code Quality | The codebase follows best practices with automated testing |
| 📄 | Documentation | Medium Article |
| 🧩 | Modularity | The codebase is modular with abstract factory modules for data loading, model creating, training and testing, even for inference for single testing |
| 🧪 | Testing | src/inference.py |
| 📦 | Dependencies | Key dependencies include Python, HuggingFace and CUML |
Model Architecture Detail
check on src/model.py
Train
$ bash src/train.sh
Test
$ python src/inference.py
- Report Issues: Submit bugs found or log feature requests.
This project is protected under the LICENSE file.