home/categories/machine-learning/benchflow-ai-skillsbench-tasks-mhc-layer-impl-environment-skills-nanogpt-training-skill-md
machine-learningdata-ai

nanogpt-training

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.

benchflow-ai
maintainer
benchflow-ai
Mis à jour 1/19/2026
Étoiles
946
Forks
244
quick start

Installation and usage

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.

Installation
$ install --globalskills.sh
Utilisation

Après l'installation, vous pouvez utiliser ce skill en exécutant la commande suivante dans votre terminal :

skills use nanogpt-training