The company’s new NVLM 1.0 family of large multimodal language models, led by the 72 billion parameter NVLM-D-72B, demonstrates exceptional performance across vision and language tasks while ...