Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/anyesh/emotion-recognition
AI-based application for emotion detection and recognition from text data
https://github.com/anyesh/emotion-recognition
machine-learning nlp nlp-machine-learning
Last synced: about 1 month ago
JSON representation
AI-based application for emotion detection and recognition from text data
- Host: GitHub
- URL: https://github.com/anyesh/emotion-recognition
- Owner: Anyesh
- Created: 2020-05-07T09:45:00.000Z (over 4 years ago)
- Default Branch: master
- Last Pushed: 2022-11-22T09:58:48.000Z (about 2 years ago)
- Last Synced: 2023-02-27T23:57:20.135Z (almost 2 years ago)
- Topics: machine-learning, nlp, nlp-machine-learning
- Language: Jupyter Notebook
- Homepage:
- Size: 1.8 MB
- Stars: 1
- Watchers: 2
- Forks: 2
- Open Issues: 4
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Emotion Detection and Recognition from Text data
## Project Structure
```
├── README.md <- README file.
├── api <- APIs to interact with the inference model.
│ ├── example.py
|
├── data
│ ├── example.csv <- raw data from third party sources.
|
├── docs <- Project related analysis and other documents
│
├── models <- Trained and serialized models/artifacts
| |── v1
| |── artifact.h5
| |── v2
| |── artifact.h5
│
├── notebooks <- Data analysis Jupyter notebooks
│
├── requirements.txt <- Pip generated requirements file for the project.
│
├── emotion_detection <- Source code for use in this project.
│ ├── __init__.py
│ │
│ ├── config <- Contains the config files.
│ │ └── config.py
| |
│ ├── data <- Scripts to download data and store on root data path.
│ │ └── make_dataset.py
| |
│ ├── dispatcher <- Collection of various ML models ready to dispatch.
│ │ └── dispatcher.py
│ │
│ ├── features <- Scripts to process the data.
│ │ └── build_features.py
│ │
│ ├── models <- Scripts to train, test, and build model
│ │ │
│ │ ├── test_model.py
│ │ └── train_model.py
│ │ └── build_model.py
| |
│ ├── utils <- Collection of various utility functions.
| | └── example.py
| |
│ ├── run_app.py <- script to run the flask web app
│ ├── run.py <- script to run the model training
│ ├── simple_inference.py <- script to test the model on cli
```
## Getting Started
### Requirements
```
pip install -r requirements.txt
```
## Config File
Config file at `emotion_detection/config/config.py` contains all the necessary configurations. Please make sure to check that before preoceeding.
### Example:
```
import os
BASE_DIR = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
DATA_PATH = os.path.join(BASE_DIR, "data", "raw")
DATASET_NAME = "ISEAR_dataset.csv"
DATASET_URL =
MODEL_PATH = os.path.join(BASE_DIR, "models")
CHECKPOINT_PATH = os.path.join(BASE_DIR, "checkpoints")
```
### Dispatcher
All the available ML models should be listed in the `emotion_detection/dispatcher/dispatcher.py` file. This will be used as the `model-name` while training and testing.
Example:
```
MODELS = {
"randomforest": ensemble.RandomForestClassifier,
"naive_bayes": MultinomialNB,
"xgboost": XGBClassifier,
"logistic": LogisticRegression,
"sgd_classifier": SGDClassifier,
"svm_svc": SVC,
}
```
### Model parameters
Hyperparameters for the listed models are to be stored in the `emotion_detection/config/model_params.py` file with the same name as the listed models in dispatcher.
Example:
```
"bert_classifier": {},
"xgboost": {},
"randomforest": {},
"naive_bayes": {"alpha": 0.1},
```
### Download the dataset
The following command will download the dataset from the URL given in `src/config/config.py` file .
```
python -m emotion_detection.data.make_dataset
```
### Train the models
```
python run.py --model-name --vocab-size --train-size
```
### Test - Simple inference
```
python simple_inference.py --model-name
```
Example:
```
python run.py --model-name naiv_bayes --vocab-size 7000 --train-size 0.7
```
## Flask Web App
To run the Flask application in docker with MongoDB run the following command.
Configure the MongoDB URL and DB name at `api/config.cfg`.
Change to local DB if not using docker.
```
python run_app.py
```
## Run in docker
```
docker-compose up
```
## Try running modules seperately
### Train the model
The following command will train the model by first pre-processing the dataset from the `feature_generator.py` and train on the configured ML model.
```
python -m emotion_detection.models.train_model
```
### Test the model
```
python -m emotion_detection.models.test_model
```
## To-do List
- [x] Download dataset
- [x] Pre-process data
- [x] Train model
- [x] Test model
- [x] Main Pipeline
- [x] Flask app
- [ ] Clean build