GPTZero

Based on BurhanUlT's repo

gpt0_demo.mp4

Still very much a work in progress.

• Line-by-line analysis of text
• Make UI more user-friendly
• Add PDF support
• Update the model to support longer context window.
• Add more explaination to what is happening. (eh)
• Add cool animations (eh)

How to run

1. Clone the repo

2. Navigate to the directory

$ cd GPTZeroWeb
...
$ docker compose build
...

Building would take around 7 minutes, because spaCy does not have a pre-bult wheel.

3. Run the container

$ docker compose up

4. Navigate to localhost:8000 in your browser.

To alter this project.

Aside from having docker and everything ready, you would want to install all the requirements in your python environment, and install the en_core_web_sm model for spaCy.

$ pip install -r requirements.txt
...
$ python -m spacy download en_core_web_sm
...

---

Copied from the original README:

GPTZero is an AI model with some mathematical formulation to determine if a particular text fed to it is written by AI or a human being.

Note

This implementation produces 100% same results as gptzero.me. We've compared extensively on a large corpus of text to compare our values with them, and surprisingly got the same results.

Acknowledgement

1. This repository is built based on the hugging face https://huggingface.co/docs/transformers/perplexity

2. Liu, Yinhan, et al. "Roberta: A robustly optimized bert pretraining approach." arXiv preprint arXiv:1907.11692 (2019).

3. BurhanUlT's repo