mudittt/text-summarizer

It is an end-to-end text summarizer application, which uses Meta's BART model and is fine-tuned on the Samsung dataset.

Clone the repo

git clone https://github.com/mudittt/text-summarizer.git

Setup a conda environment

conda create --name torch python=3.8
conda activate torch

• if you're using a macbook with an apple chip, then install the nightly version of pytorch to use pytorch with gpu support on your mac.

• install all other remaining package from requirements.txt

conda install pytorch-nightly::pytorch torchvision torchaudio -c pytorch-nightly

conda install transformers accelerate sentencepiece

conda install sacrebleu rouge-score py7zr pandas boto3 mypy-boto3-s3 nltk matplotlib notebook uvicorn fastapi python-box 

conda install protobuf

Workflows

- Update config.yaml
- Update params.yaml
- Update entity
- Update the configuration manager in src config
- update the conponents
- update the pipeline
- update the main.py
- update the app.py