ai-trends

This repo contains a prototype for an AI trends summarizer. The code can be used to summarize the abstracts of the top AI papers per month according to HuggingFace daily papers. It also creates a list of the top concepts for each month.

Requirements

• Python >= 3.7
• Additional packages are defined in the requirements.txt file.
• Token for HuggingFace Serverless Inference, a free account is fine.

It is recommended to create a virtual environment and install the packages there.

python -m venv .venv
source .venv/bin/activate
python -m pip install -r requirements.txt

Running

First, clone the repo. Then, run the summarizer as follows.

python summarize_hf.py -m '01,02' -y 2025 -k 5 -o hf_report.html -t <HUGGINGFACE_TOKEN>

Options

• -m, months to download, note the leading 0 for single digit month numbers, and use quotes for multiple months, eg. '01,02,04'.
• -y, years to download, same syntax as months.
• -k, How many papers to summarize for each month.
• -o, filename to write the html report to.
• -t, HuggingFace token with Serverless Inference access.