New: gpt-oss is an open leading reasoning language model designed to fit onto a single machine.
I study Artificial Intelligence. Check out my published work:
- GPT-2 Metadata Pretraining Towards Instruction Finetuning for Ukrainian (UNLP 2023 with Dmytro Chaplynskyi). https://github.com/proger/uk4b
- Re: VAE Approximation Error: ELBO and Exponential Families (ReScience C, Tweet, NeurIPS 2023 Journal Track Poster with Navdeep Singh Bedi and Qianbo Zang). https://github.com/proger/vae
- Setting up the Data Printer with Improved English to Ukrainian Machine Translation (UNLP 2024 with Yurii Paniv, Dmytro Chaplynskyi, Nikita Trynus). arXiv Model
Projects:
pip install accelerated-scan: Accelerated Scan is a high-performance GPU implementation of the first order associative scan for Torch.pip install world-model: Learning from simulation by predicting the next frame and action: https://github.com/proger/world-modelpip install haloop: a speech recognition and language modelling toolkit for noisy data
I write notes in my blog, https://proger.github.io. Check out my notes on Bayesian Flow Networks
