Metadata-Version: 2.1
Name: EMO-AI
Version: 0.0.3
Summary: library for the ai competition, currently private
Home-page: https://github.com/Kelvinthedrugger/EMO_AI/tree/master/
Author: Kelvinthedrugger
Author-email: aangus0628@gmail.com
License: Apache Software License 2.0
Keywords: emotional pointer
Platform: UNKNOWN
Classifier: Development Status :: 3 - Alpha
Classifier: Intended Audience :: Developers
Classifier: Natural Language :: English
Classifier: Programming Language :: Python :: 3.7
Classifier: Programming Language :: Python :: 3.8
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: License :: OSI Approved :: Apache Software License
Requires-Python: >=3.7
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: pip
Requires-Dist: packaging
Requires-Dist: tokenizers (>=0.12.1)
Requires-Dist: transformers (>=4.20.1)
Requires-Dist: torch (>=1.12.0)
Provides-Extra: dev

# EMO_AI
Use state-of-the-art to detect the user's emotion on social apps, particularly needed in modern society

## Installation
#### it's recommended to install pytorch via [official guide](https://pytorch.org/) first

    # stable version: have to install transformers, tokenizers, torch by hand ...
    pip install EMO-AI==0.0.2

    # latest version
    pip install EMO-AI


## Usage

    from EMO_AI.model_api import *
    from EMO_AI.data_process import *
    t = "Elvis is the king of rock"
    tokenizer = get_tokenizer()
    PATH = "your_pretrained_model.pt"
    # check how the model is saved in the first place
    model = get_model(PATH, inference_only=True)
    import torch
    with torch.no_grad():
            model.eval() # evaluate mode
            # convert_text_to_tensor(t) would work, but kinda slower and wasteful
            rep = model(convert_text_to_tensor(t, tokenizer))
    # print output tensor from forward pass
    print(rep)
    # get predicted emotion
    print_emotion(rep)



