ParVL: Parallel Scaling and Expandable Compute Allocation for Multimodal LLMs
- Published
- Source
- arXiv
- Paper number
- 816
- Field
- Computer Vision
- arXiv ID
- 2608.04010
Key points
- It proposes the first framework that creates separate parallel branches for ViT and LLM so that compute allocation can be adjusted independently.
- A prefix-conditioned branch shares backbone parameters, so compute can be expanded without increasing parameters.
- Systematic experiments on nine vision-language branch configurations show that the optimal allocation differs by task.
- It confirms consistent gains over a single-branch model at 1B, 2B, and 8B scale.
- It analyzes that parallel execution does not directly translate extra compute into wall-clock latency.
Paper links
External research summaries. These are not HDATF publications or measured product results.