Metadata-Version: 2.1
Name: arxiv_agent
Version: 0.0.0
Summary: LLM agent to answer questions about arXiv papers
Description-Content-Type: text/markdown
Requires-Dist: numpy
Requires-Dist: langchain
Requires-Dist: langgraph
Requires-Dist: langchain-openai
Requires-Dist: langchain-community
Requires-Dist: termcolor
Requires-Dist: bs4
Requires-Dist: arxiv
Requires-Dist: semanticscholar

# arXiv agent

## Setup
```
conda create -n arxiv-agent python=3.12
conda activate arxiv-agent
# pytorch packages
conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia
# huggingface packages
pip install transformers accelerate
```

```
conda env config vars set LANGCHAIN_TRACING_V2="true"
conda env config vars set LANGCHAIN_API_KEY=[...]
```

Install `arxiv_agent` package:
```
conda activate arxiv-agent
conda develop .
```

CodeLLama:


## Design

Prompt:
1. Question
2. Available API calls: you can use the following models
- Specification of what each model does
For now, I'm thinking we can just insert "specific" knowledge into each model via context
- 
3. State

## Data

Small dataset:


Medium dataset:
https://huggingface.co/datasets/natural_questions
