Where Is VLA Inference Time Spent? Bottlenecks and Optimization Interactions Across GPUs
cs.RO
Submitted: 2026-11-04
Updated: 2026-11-04
Code: https://github.com/flashinfer-ai/flashinfer
Terminology
Sources
- ActQuant: Sub-4-bit Action-Guided Backbone Weight-Only Quantization for Vision-Language-Action Models
- GR00T N1: An Open Foundation Model for Generalist Humanoid Robots
- VLA-IAP: Training-Free Visual Token Pruning via Interaction Alignment for Vision-Language-Action Models
- SQAP-VLA: A Synergistic Quantization-Aware Pruning Framework for High-Performance Vision-Language-Action Models
- $\pi_{0.5}$: a Vision-Language-Action Model with Open-World Generalization
- How Fast Can I Run My VLA? Demystifying VLA Inference Performance with VLA-Perf
- Beyond LLM Serving: Characterizing Vision-Language-Action Workloads for Embodied AI System Design
- OpenVLA: An Open-Source Vision-Language-Action Model
- Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success
- CogACT: A Foundational Vision-Language-Action Model for Synergizing Cognition and Action in Robotic Manipulation
- DepthCache: Depth-Guided Training-Free Visual Token Merging for Vision-Language-Action Model Inference
- Flow Matching for Generative Modeling
- Bridging the Semantic-Action Gap in Visual Token Pruning for Efficient VLA Inference
- SnapFlow: One-Step Action Generation for Flow-Matching VLAs via Progressive Self-Distillation
- Quantization-Aware Imitation-Learning for Resource-Efficient Robotic Control
- SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Robotics
- Denoising Diffusion Implicit Models
- Rethinking the Practicality of Vision-language-action Model: A Comprehensive Benchmark and An Improved Baseline
- Think Twice, Act Once: Token-Aware Compression and Action Reuse for Efficient Inference in Vision-Language-Action Models
- Fewer Steps, Better Actions: Rethinking Flow-Matching Inference for VLA Policies
Related papers
- Rethinking Actions: Continuity-Aware Priors for World-Model Planning
- MOSAIC: Modular Obstacle and Spatial-Aware Framework for Manipulation In Cluttered Scenes
- SAGAS: Semantic-Aware Graph-Assisted Stitching for Offline Temporal Logic Planning
- DexSeed: Expanding Human Hand-Object Interactions into Robot Demonstrations for Dexterous Policy Learning
- LaST0.5: A Generalist World Action Model with Interaction-Centric Latent Reasoning
- CIALLO: Learning Shared Intent Representations for Cross-Embodiment Manipulation