Search results for NeMo Curator
AI & Enterprise
Nvidia-backed Korean sovereign AI developers use Nvidia technology from training to inference
Developers of government-backed Korean sovereign AI foundation models shared how they use Nvidia platforms and software across data processing, model training and inference. At an event in Seoul, they highlighted Nvidia\'s Dynamo inference optimisation software, NeMo Curator for removing duplicate training documents and Megatron Bridge for training and conversion. Nvidia also unveiled Dynamo optimisations for four Korean AI models and cited test results showing higher throughput.
Telecommunications & Media
SKT steps up scale-up of proprietary AI to multimodal and agents
SK Telecom has passed a second-stage evaluation of the government’s project for a proprietary AI foundation model. The assessment reflected industrial applicability and impact on South Korea’s AI ecosystem. SKT plans to expand its technology scope in the next evaluation to models with up to trillion-level parameters and to multimodal and AI agent areas. It has developed the A.X K1 and expanded to A.X K2, and it is building derived vision and speech models and applying them in manufacturing and defence.
AI & Enterprise
Motif builds proprietary tools instead of Nvidia software for homegrown AI foundation model development
Motif Technologies, a developer of proprietary AI foundation models, is seeking differentiation by using its own platform rather than Nvidia software tools. It says using Nvidia’s general-purpose tools makes it hard to beat big tech firms with superior computing power. Chief Executive Jung-hwan Lim said Nvidia tools limit developer flexibility and can lock users into Nvidia’s ecosystem. Motif uses Nvidia GPUs and CUDA but replaces Nvidia’s model development tools with in-house software.
