{"type": "article", "title": "Extending Raschka's GPT-2: an MoE trained from scratch on an RTX 3090", "publisher": "Web Pulse", "url": "https://wpnews.pro/news/extending-raschka-s-gpt-2-an-moe-trained-from-scratch-on-an-rtx-3090", "original_source": "https://www.gilesthomas.com/2026/09/gpt-2-to-moe", "published": "2026-09-10T18:45:00+00:00", "accessed": "2026-09-10", "id": "extending-raschka-s-gpt-2-an-moe-trained-from-scratch-on-an-rtx-3090"}