Reproduce “PhoGPT: Generative Pre-training for Vietnamese” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Dat Quoc Nguyen, Linh The Nguyen, Chi Tran, Dung Ngoc Nguyen, Dinh Phung, Hung Bui
PhoGPT is an open Vietnamese language-model release with a 3.7-billion-parameter base model trained from scratch on about 102 billion Vietnamese tokens, then fine-tuned for instruction and conversation. On ViTruthfulQA, PhoGPT-4B-Chat scored 43.5% on 147 Vietnam-specific questions, ahead of tested peers, but GPT-4 scored 44.7% overall, limiting the claim to this benchmark.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.