We're thrilled to release Holotron-12B, a multimodal computer-use model from H Company. Post-trained from the open NVIDIA Nemotron-Nano-2 VL model on H Company’s proprietary data mixture, Holotron-12B is the result of a close collaboration between our research labs to engineer a new type of model optimized primarily for scale and performance in production.

Most multimodal models today optimize primarily for static vision or following instructions. Holotron-12B, just like our Holo2 model, however, has a different goal: serving as a policy model for computer-use agents that must perceive, decide, and act efficiently in interactive environments.

With Holotron-12B, we wanted to create a model that could efficiently and effectively scale in production while handling long contexts with multiple images, and still perform well on agent benchmarks. The NVIDIA Nemotron model offered a strong foundation on the inference side, and by developing Holotron-12B we've demonstrated how much more the model can accomplish with further training.

Holotron-12B's significant leap in inference efficiency is made possible by its foundational Nemotron architecture, which utilizes a hybrid State-Space Model (SSM) and attention mechanism. Unlike purely transformer-based models, this design is optimized for high-throughput serving. State-space models offer superior scalability for long-context inference by avoiding the quadratic computation cost associated with the full attention mechanism, particularly benefiting agentic workloads involving multiple images and lengthy interaction histories. In terms of inference, the main contribution of an SSM is its dramatically reduced memory-footprint: while vanilla attention stores K and V activations per token and layer (the notorious KV Cache), SSMs are a linear recurrent model, storing only a constant state per layer per generated sequence, independent of the length of the sequence.

Source link