Simply specify of the paper your want to understand and obtain a short summary (TL;DR) of it in seconds.
# -----------------------------------------------------------------------------
# Setting up the environment
# Create the PaperGPT environment
conda env create -f environment.yml
# (Optiona) Run the following to check if PaperGPT got successfully created
conda env list
# Activate the environment
conda activate PaperGPT
# -----------------------------------------------------------------------------
conda install -c huggingface transformers
pip install PyPDF2
conda install -c conda-forge pypdf2
pip install openai
conda install -c conda-forge openai -y
- Ease of use: ensure that library is easy to install. Clear documentation and examples.
- Automatic processing: automatically extract text and relevant content from papers, eliminate need for users to manually input or format the content
- Compatibility: should be able to handle various document formats (
PDF
,PNG
,JPG
), and possibly even accept URLs to academic papers hosted online - Multi-language support
- Use cases: Provide real-world use cases and scenarios where the library can be valuable, such as research paper review, studying, or staying updated on the latest research.
- Model Options: Provide option for the user to decide between different LLM models such as
LLaMa
,GPT
,Bert
BART
(Facebook AI) --BERT
(Google AI) --bert-base-uncased
GPT-2
(OpenAI)