C-RADIO
An open-source vision encoder released by NVIDIA, used for image and video feature extraction in multimodal models.
Timeline (last 90 days)
Nemotron-3.5-Super-VL comes with a 1.3GB vision tower (C-RADIO) that needs to be streamed layer by layer to the GPU, taking 0.4-1.6 seconds to process each image.
Nemotron 3.5 Super VL uses a Mamba-2/Transformer hybrid MoE architecture, with 512 routed experts per MoE layer, paired on the vision side with a C-RADIO encoder, supporting text, JPEG/PNG images, and MP4 video up to 2 minutes long, with a context limit of 1M tokens and native support for multi-token prediction.
All Facts (2)
Nemotron-3.5-Super-VL comes with a 1.3GB vision tower (C-RADIO) that needs to be streamed layer by layer to the GPU, taking 0.4-1.6 seconds to process each image.
50%UnverifiedNemotron 3.5 Super VL uses a Mamba-2/Transformer hybrid MoE architecture, with 512 routed experts per MoE layer, paired on the vision side with a C-RADIO encoder, supporting text, JPEG/PNG images, and MP4 video up to 2 minutes long, with a context limit of 1M tokens and native support for multi-token prediction.
50%