Text model

Megatron-Turing NLG

Megatron-Turing NLG is a 530B-parameter Transformer co-developed by NVIDIA and Microsoft. Trained on a large curated text corpus, it delivers strong performance in open-ended generation, long-form QA, summarization, and few-shot learning, and it helped pioneer large-scale training with Megatron and DeepSpeed.

Type
Text
Weights
Open