mlsanityData to 7 Oct 2026

FoldQuantVLA: Native Low-Bit Quantization of Vision-Language-Action Models via Consistent Folding

arXiv 2609.24433 ↗First posted 21 Sep 2026cs.RO cs.AI cs.SYcited by 0

On the map: Vision-language-action models → Efficient VLA inference and adaptation

Abstract (arXiv)

Low-bit vision-language-action inference must reduce observation-to-action latency while preserving robot behavior. We present FoldQuantVLA, a post-training quantization framework that carries a consistent activation representation through calibration, weight rounding, and native integer execution. It combines channel scaling and block Hadamard transforms with dynamic per-token quantization, without policy retraining. Custom TensorRT plugins execute projections in both the language backbone and iterative action expert with four-bit weights and activations (W4A4) on Ada GPUs and Jetson AGX Orin. Evaluation spans LIBERO, SimplerEnv, and two robot platforms. Across three GR00T checkpoints and π0.5, W4A4 achieves 1.20 to 1.33× speedups over floating-point TensorRT on Orin and 1.25 to 1.52× on desktop. Retaining language attention-output and feed-forward down projections at eight bits (W8A8) improves held-out action fidelity on all four checkpoints. Across four real-robot tasks, this configuration raises observed GR00T N1.7 success from 80.0% with uniform W4A4 to 92.5% over 80 trials per configuration, with a measured additional Orin latency of 1 ms.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Efficient VLA inference and adaptation, one of 13 directions in Vision-language-action models. Builds on 31 of its 36 references that are in the map. First posted on arXiv 21 Sep 2026.

Builds on

31 of 36 references.

Closest ancestors · 5

Same direction, newest first

  1. 2026HoloQ-VLA: Uniform W4A4 Quantization of Vision-Language-Action Models27 May
  2. 2026ActQuant: Sub-4-bit Action-Guided Quantization for Vision-Language-Action Models19 May
  3. 2026DyQ-VLA: Temporal-Dynamic-Aware Quantization for Embodied Vision-Language-Action Models9 Mar
  4. 2026QuantVLA: Scale-Calibrated Post-Training Quantization for Vision-Language-Action Models23 Feb
  5. 2026QVLA: Not All Channels Are Equal in Vision-Language-Action Model's Quantization3 Feb

Other references · 26

  1. 2022GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformerscited by 1,838
  2. 2025π0.5: a Vision-Language-Action Model with Open-World Generalizationcited by 1,653
  3. 2022SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Modelscited by 1,338
  4. 2023LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learningcited by 1,338
  5. 2023AWQ: Activation-aware Weight Quantization for LLM Compression and Accelerationcited by 1,234
  6. 2025GR00T N1: An Open Foundation Model for Generalist Humanoid Robotscited by 1,167
  7. 2024QuaRot: Outlier-Free 4-Bit Inference in Rotated LLMscited by 598
  8. 2025SmolVLA: A Vision-Language-Action Model for Affordable and Efficient Roboticscited by 507
  9. 2024Evaluating Real-World Robot Manipulation Policies in Simulationcited by 484
  10. 2024SpinQuant: LLM quantization with learned rotationscited by 446
  11. 2024SliceGPT: Compress Large Language Models by Deleting Rows and Columnscited by 378
  12. 2023Atom: Low-bit Quantization for Efficient and Accurate LLM Servingcited by 255
14 more
  1. 2024QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Servingcited by 200
  2. 2024DuQuant: Distributing Outliers via Dual Transformation Makes Stronger Quantized LLMscited by 165
  3. 2024SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Modelscited by 161
  4. 2024FlatQuant: Flatness Matters for LLM Quantizationcited by 116
  5. 2025OstQuant: Refining Large Language Model Quantization with Orthogonal and Scaling Transformations for Better Distribution Fittingcited by 94
  6. 2025GPTAQ: Efficient Finetuning-Free Quantization for Asymmetric Calibrationcited by 54
  7. 2025Evo-1: Lightweight Vision-Language-Action Model with Preserved Semantic Alignmentcited by 35
  8. 2025Qronos: Correcting the Past by Shaping the Future... in Post-Training Quantizationcited by 24
  9. 2024MixLLM: LLM Quantization with Global Mixed-precision between Output-features and Highly-efficient System Designcited by 17
  10. 2024BitNet a4.8: 4-bit Activations for 1-bit LLMscited by 16
  11. 2024HadaCore: Tensor Core Accelerated Hadamard Transform Kernelcited by 9
  12. 2025SmoothRot: Combining Channel-Wise Scaling and Rotation for Quantization-Friendly LLMscited by 3
  13. 2025AccuQuant: Simulating Multiple Denoising Steps for Quantizing Diffusion Modelscited by 3
  14. 2026Q-Drift: Quantization-Aware Drift Correction for Diffusion Model Samplingcited by 3

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026HBVLA: Pushing 1-Bit Post-Training Quantization for Vision-Language-Action Models
  2. 2025FASTer: Toward Efficient Autoregressive Vision Language Action Modeling via Neural Action Tokenization
  3. 2025BitVLA: 1-bit Vision-Language-Action Models for Robotics Manipulation
  4. 2025Think Twice, Act Once: Token-Aware Compression and Action Reuse for Efficient Inference in Vision-Language-Action Models
  5. 2026A1: A Fully Transparent Open-Source, Adaptive and Efficient Truncated Vision-Language-Action Model
  6. 2026CHASE-VLA: Post-Training Quantization Framework for Vision-Language-Action Models with Chunk-Aware Scale Estimation
  7. 2026VLAQuantBench: Closed-Loop Evaluation of Post-Training Quantization for Vision-Language-Action Models
  8. 2026Finetuning Vision-Language-Action Models Requires Fewer Layers Than You Think