nanogpt-training by benchflow-ai

nanogpt-training

benchflow-ai

9392423 months ago

Train GPT-2 scale models (~124M parameters) efficiently on a single GPU. Covers GPT-124M architecture, tokenized dataset loading (e.g., HuggingFace Hub shards), modern optimizers (Muon, AdamW), mixed precision training, and training loop implementation.

>_
Quick Install
npxskills add benchflow-ai/skillsbench--skill nanogpt-training
Instructions

No instructions available