VLA language backbones show high redundancy on manipulation benchmarks, with half the LLM blocks removable and even two blocks sufficient to recover baseline performance after fine-tuning, unlike vision and action pathways.
super hub Mixed citations
Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success
Mixed citation behavior. Most common role is background (64%).
178
Pith papers citing it
Background
64% of classified citations
abstract
Recent vision-language-action models (VLAs) build upon pretrained vision-language models and leverage diverse robot datasets to demonstrate strong task execution, language following ability, and semantic generalization. Despite these successes, VLAs struggle with novel robot setups and require fine-tuning to achieve good performance, yet how to most effectively fine-tune them is unclear given many possible strategies. In this work, we study key VLA adaptation design choices such as different action decoding schemes, action representations, and learning objectives for fine-tuning, using OpenVLA as our representative base model. Our empirical analysis informs an Optimized Fine-Tuning (OFT) recipe that integrates parallel decoding, action chunking, a continuous action representation, and a simple L1 regression-based learning objective to altogether improve inference efficiency, policy performance, and flexibility in the model's input-output specifications. We propose OpenVLA-OFT, an instantiation of this recipe, which sets a new state of the art on the LIBERO simulation benchmark, significantly boosting OpenVLA's average success rate across four task suites from 76.5% to 97.1% while increasing action generation throughput by 26$\times$. In real-world evaluations, our fine-tuning recipe enables OpenVLA to successfully execute dexterous, high-frequency control tasks on a bimanual ALOHA robot and outperform other VLAs ($\pi_0$ and RDT-1B) fine-tuned using their default recipes, as well as strong imitation learning policies trained from scratch (Diffusion Policy and ACT) by up to 15% (absolute) in average success rate. We release code for OFT and pretrained model checkpoints at https://openvla-oft.github.io/.
hub tools
citation-role summary
background 34
baseline 10
method 6
other 3
citation-polarity summary
claims ledger
- abstract Recent vision-language-action models (VLAs) build upon pretrained vision-language models and leverage diverse robot datasets to demonstrate strong task execution, language following ability, and semantic generalization. Despite these successes, VLAs struggle with novel robot setups and require fine-tuning to achieve good performance, yet how to most effectively fine-tune them is unclear given many possible strategies. In this work, we study key VLA adaptation design choices such as different action decoding schemes, action representations, and learning objectives for fine-tuning, using OpenVLA
authors
co-cited works
representative citing papers
LIBERO-Safety supplies a scalable benchmark, data-generation pipeline, and 19,664-demonstration dataset that exposes a generalization-safety tension in current VLA models where diverse training improves collision avoidance but task success stays limited by trajectory quality and semantic understandi
DVAC uses denoising variance as an intrinsic signal to adaptively chunk actions in flow-based robot policies, improving success rates and cutting replans on LIBERO, RoboTwin, CALVIN, and real-world tasks.
BOKBO is the first conformal abstention method for K-sample VLA policies that supplies finite-sample distribution-free guarantees on executed violation rates, with global and Mondrian per-task variants.
JOPAT jointly models pixels, point tracks, and actions in a diffusion transformer and reports gains over pixel-only baselines on long-horizon robot tasks with occlusion and off-screen motion.
Pion modifies Muon's Newton-Schulz iterations into a controllable high-pass filter that anchors dominant singular values at 1 while suppressing noisy tails, outperforming Muon and AdamW in VLA and RLVR regimes.
PCM uses success-failure action variance to probabilistically select and mask chunks for gradient updates in GRPO, matching standard success rates with 2.38x wall-clock speedup and 60% lower memory on LIBERO benchmarks.
DSSP is a history-conditioned diffusion state space policy that uses SSMs to encode full observation streams with an auxiliary dynamics objective and hierarchical fusion, achieving SOTA results with reduced model size in robot manipulation.
RotVLA models latent actions as continuous SO(n) rotations with triplet-frame supervision and flow-matching to reach 98.2% success on LIBERO and 89.6%/88.5% on RoboTwin2.0 using a 1.7B-parameter model.
GridS is a plug-and-play differentiable module for geometry-aware visual token resampling in VLA models that achieves under 10% token retention and 76% FLOPs reduction with no success-rate loss.
MCF-Proto adds a motion-centric local action frame and prototype parameterization to VLA models, inducing emergent geometric structure and improved robustness from standard demonstrations alone.
Capability vectors extracted from parameter differences between standard and auxiliary-finetuned VLA models can be merged into pretrained weights to match auxiliary-training performance while reducing computational overhead during adaptation.
VEGA improves spatial reasoning in VLA models for robotics by aligning visual encoder features with 3D-supervised DINOv2 representations via a temporary projector and cosine similarity loss.
GuardVLA embeds a stealthy backdoor watermark in VLAs via secret messages in visual data and uses a swap-and-detect mechanism for post-release ownership verification that preserves task performance.
Reducing visual input to one token per frame in VLA world models maintains or improves long-horizon performance on MetaWorld, LIBERO, and real-robot tasks.
OA-WAM uses persistent address vectors and dynamic content vectors in object slots to enable addressable world-action prediction, improving robustness on manipulation benchmarks under scene changes.
CoRAL lets LLMs act as adaptive cost designers for motion planners while using VLM priors and online identification to handle unknown physics, achieving over 50% higher success rates than baselines in unseen contact-rich robotic scenarios.
VUDA enables spatial sharing between CUDA and Vulkan on GPUs via channel redirection and page-table grafting, achieving up to 85% higher throughput than temporal baselines in embodied AI tasks.
A multimodal transformer generates and caches interleaved text-image traces to guide closed-loop actions, achieving 92.4% success on LIBERO-Long and 95.5% average on LIBERO.
Introduces ISS and NMR as interventional metrics to diagnose causal misalignment in VLA policies and link it to generalization performance.
Being-H0.7 adds future-aware latent reasoning to direct VLA policies via dual-branch alignment on latent queries, matching world-model benefits at VLA efficiency.
Discrete diffusion policies act as natural asynchronous executors for robotics by treating action generation as iterative unmasking, yielding higher success rates and lower computation than flow-matching real-time chunking in dynamic tasks.
Libra-VLA introduces a coarse-to-fine dual-system architecture for VLA models that decouples discrete macro-directional planning from continuous micro-pose refinement, with performance peaking at balanced learning difficulty.
SpecRLBench is a new benchmark evaluating generalization of LTL-guided RL methods across navigation and manipulation domains with static/dynamic environments and varied robot dynamics.
citing papers explorer
-
Drop-Then-Recovery: How Redundant Are Vision-Language-Action Models?
VLA language backbones show high redundancy on manipulation benchmarks, with half the LLM blocks removable and even two blocks sufficient to recover baseline performance after fine-tuning, unlike vision and action pathways.
-
LIBERO-Safety: A Comprehensive Benchmark for Physical and Semantic Safety in Vision-Language-Action Models
LIBERO-Safety supplies a scalable benchmark, data-generation pipeline, and 19,664-demonstration dataset that exposes a generalization-safety tension in current VLA models where diverse training improves collision avoidance but task success stays limited by trajectory quality and semantic understandi
-
Denoising Tells When to Replan: Denoising-Variance Adaptive Chunking for Flow-Based Robot Policies
DVAC uses denoising variance as an intrinsic signal to adaptively chunk actions in flow-based robot policies, improving success rates and cutting replans on LIBERO, RoboTwin, CALVIN, and real-world tasks.
-
BOKBO (Best of K Bad Options): Calibrated Abstention for VLA Policies
BOKBO is the first conformal abstention method for K-sample VLA policies that supplies finite-sample distribution-free guarantees on executed violation rates, with global and Mondrian per-task variants.
-
Point Tracking Improves World Action Models
JOPAT jointly models pixels, point tracks, and actions in a diffusion transformer and reports gains over pixel-only baselines on long-horizon robot tasks with occlusion and off-screen motion.
-
Rethinking Muon Beyond Pretraining: Spectral Failures and High-Pass Remedies for VLA and RLVR
Pion modifies Muon's Newton-Schulz iterations into a controllable high-pass filter that anchors dominant singular values at 1 while suppressing noisy tails, outperforming Muon and AdamW in VLA and RLVR regimes.
-
Learn Where Outcomes Diverge: Efficient VLA RL via Probabilistic Chunk Masking
PCM uses success-failure action variance to probabilistically select and mask chunks for gradient updates in GRPO, matching standard success rates with 2.38x wall-clock speedup and 60% lower memory on LIBERO benchmarks.
-
DSSP: Diffusion State Space Policy with Full-History Encoding
DSSP is a history-conditioned diffusion state space policy that uses SSMs to encode full observation streams with an auxiliary dynamics objective and hierarchical fusion, achieving SOTA results with reduced model size in robot manipulation.
-
RotVLA: Rotational Latent Action for Vision-Language-Action Model
RotVLA models latent actions as continuous SO(n) rotations with triplet-frame supervision and flow-matching to reach 98.2% success on LIBERO and 89.6%/88.5% on RoboTwin2.0 using a 1.7B-parameter model.
-
See What Matters: Differentiable Grid Sample Pruning for Generalizable Vision-Language-Action Model
GridS is a plug-and-play differentiable module for geometry-aware visual token resampling in VLA models that achieves under 10% token retention and 76% FLOPs reduction with no success-rate loss.
-
Beyond World-Frame Action Heads: Motion-Centric Action Frames for Vision-Language-Action Models
MCF-Proto adds a motion-centric local action frame and prototype parameterization to VLA models, inducing emergent geometric structure and improved robustness from standard demonstrations alone.
-
CapVector: Learning Transferable Capability Vectors in Parametric Space for Vision-Language-Action Models
Capability vectors extracted from parameter differences between standard and auxiliary-finetuned VLA models can be merged into pretrained weights to match auxiliary-training performance while reducing computational overhead during adaptation.
-
VEGA: Visual Encoder Grounding Alignment for Spatially-Aware Vision-Language-Action Models
VEGA improves spatial reasoning in VLA models for robotics by aligning visual encoder features with 3D-supervised DINOv2 representations via a temporary projector and cosine similarity loss.
-
Towards Backdoor-Based Ownership Verification for Vision-Language-Action Models
GuardVLA embeds a stealthy backdoor watermark in VLAs via secret messages in visual data and uses a swap-and-detect mechanism for post-release ownership verification that preserves task performance.
-
One Token Per Frame: Reconsidering Visual Bandwidth in World Models for VLA Policy
Reducing visual input to one token per frame in VLA world models maintains or improves long-horizon performance on MetaWorld, LIBERO, and real-robot tasks.
-
OA-WAM: Object-Addressable World Action Model for Robust Robot Manipulation
OA-WAM uses persistent address vectors and dynamic content vectors in object slots to enable addressable world-action prediction, improving robustness on manipulation benchmarks under scene changes.
-
CoRAL: Contact-Rich Adaptive LLM-based Control for Robotic Manipulation
CoRAL lets LLMs act as adaptive cost designers for motion planners while using VLM priors and online identification to handle unknown physics, achieving over 50% higher success rates than baselines in unseen contact-rich robotic scenarios.
-
VUDA: Breaking CUDA-Vulkan Isolation for Spatial Sharing of Compute and Graphics on the Same GPU
VUDA enables spatial sharing between CUDA and Vulkan on GPUs via channel redirection and page-table grafting, achieving up to 85% higher throughput than temporal baselines in embodied AI tasks.
-
Thinking in Text and Images: Interleaved Vision--Language Reasoning Traces for Long-Horizon Robot Manipulation
A multimodal transformer generates and caches interleaved text-image traces to guide closed-loop actions, achieving 92.4% success on LIBERO-Long and 95.5% average on LIBERO.
-
Embodied Interpretability: Linking Causal Understanding to Generalization in Vision-Language-Action Models
Introduces ISS and NMR as interventional metrics to diagnose causal misalignment in VLA policies and link it to generalization performance.
-
Being-H0.7: A Latent World-Action Model from Egocentric Videos
Being-H0.7 adds future-aware latent reasoning to direct VLA policies via dual-branch alignment on latent queries, matching world-model benefits at VLA efficiency.
-
DiscreteRTC: Discrete Diffusion Policies are Natural Asynchronous Executors
Discrete diffusion policies act as natural asynchronous executors for robotics by treating action generation as iterative unmasking, yielding higher success rates and lower computation than flow-matching real-time chunking in dynamic tasks.
-
Libra-VLA: Achieving Learning Equilibrium via Asynchronous Coarse-to-Fine Dual-System
Libra-VLA introduces a coarse-to-fine dual-system architecture for VLA models that decouples discrete macro-directional planning from continuous micro-pose refinement, with performance peaking at balanced learning difficulty.
-
SpecRLBench: A Benchmark for Generalization in Specification-Guided Reinforcement Learning
SpecRLBench is a new benchmark evaluating generalization of LTL-guided RL methods across navigation and manipulation domains with static/dynamic environments and varied robot dynamics.
-
CF-VLA: Efficient Coarse-to-Fine Action Generation for Vision-Language-Action Policies
CF-VLA uses a coarse initialization over endpoint velocity followed by single-step refinement to achieve strong performance with low inference steps on CALVIN, LIBERO, and real-robot tasks.
-
Mini-BEHAVIOR-Gran: Revealing U-Shaped Effects of Instruction Granularity on Language-Guided Embodied Agents
Mini-BEHAVIOR-Gran benchmark reveals a U-shaped effect of instruction granularity on embodied agent performance, with planning-width correlating best and coarse instructions linked to vision-dominant shallow policies.
-
HazardArena: Evaluating Semantic Safety in Vision-Language-Action Models
HazardArena shows VLA models trained on safe data frequently produce unsafe actions in semantically risky but visually similar settings, and a training-free Safety Option Layer reduces those failures with little performance cost.
-
STRONG-VLA: Decoupled Robustness Learning for Vision-Language-Action Models under Multimodal Perturbations
STRONG-VLA uses decoupled two-stage training to improve VLA model robustness, yielding up to 16% higher task success rates under seen and unseen perturbations on the LIBERO benchmark.
-
CT-1: Vision-Language-Camera Models Transfer Spatial Reasoning Knowledge to Camera-Controllable Video Generation
CT-1 transfers spatial reasoning from vision-language models to estimate camera trajectories, which are then used in a video diffusion model with wavelet regularization to produce controllable videos, claiming 25.7% better accuracy than prior methods.
-
BiCoord: A Bimanual Manipulation Benchmark towards Long-Horizon Spatial-Temporal Coordination
BiCoord is a new benchmark for long-horizon tightly coordinated bimanual manipulation that includes quantitative metrics and shows existing policies like DP, RDT, Pi0 and OpenVLA-OFT struggle on such tasks.
-
VP-VLA: Visual Prompting as an Interface for Vision-Language-Action Models
VP-VLA decouples high-level reasoning from low-level control in VLA models by rendering spatial anchors as visual prompts directly in the RGB observation space, outperforming end-to-end baselines.
-
Generative Control as Optimization: Time Unconditional Flow Matching for Adaptive and Robust Robotic Control
GeCO replaces time-dependent flow matching with time-unconditional optimization, enabling adaptive inference and intrinsic OOD detection for robotic imitation learning.
-
Learning Physics from Pretrained Video Models: A Multimodal Continuous and Sequential World Interaction Models for Robotic Manipulation
PhysGen uses video models to learn physics for robots, outperforming baselines by up to 13.8% on Libero and matching specialized models in real-world tasks.
-
RL-VLA$^3$: A Flexible and Asynchronous Reinforcement Learning Framework for VLA Training
RL-VLA³ is an asynchronous RL framework for VLA training that delivers up to 85.2% higher throughput than synchronous baselines while preserving identical sample efficiency and scaling to 256 GPUs.
-
ROSA: A Robotics Foundation Model Serving System for Robot Factories
ROSA introduces shared GPU-pool serving, robotics-aware abstractions for multi-model pipelines, and factory-productivity scheduling that improves output by up to 12.06x over dedicated per-robot systems.
-
Human-Centric Transferable Tactile Pre-Training for Dexterous Robotic Manipulation
Introduces H-Tac human tactile-action dataset and TTP pre-training that unifies spaces and predicts future tactile signals to improve robotic dexterous manipulation transfer.
-
Human-as-Humanoid: Enabling Zero-Shot Humanoid Learning from Ego-Exo Human Videos with Human-Aligned Embodiments
Human-as-Humanoid converts ego-exo human videos into executable 60-DoF humanoid actions through embodiment alignment and retargeting, enabling zero-shot real-robot policy deployment without target-task teleoperation data.
-
MIRTH: Mutual-Information Reasoning with Temporal Hubs for Vision-Language-Action Agents
MIRTH introduces dual-scale temporal hubs, MI-based latent reasoning tokens, and parallel decoding to VLA backbones, claiming SOTA results and error recovery on LIBERO and real LeRobot setups.
-
SA-VLA: State-aware tokenizer for improving Vision-Language-Action Models' performance
SA-VLA adds state conditioning to VQ-based action tokenization in VLA policies, expanding each discrete token's effective support to state-dependent actions and raising average success rates from 0.29 to 0.56 on 12 sim tasks and 0.15 to 0.33 on 3 real tasks.
-
Trust Your Instincts: Confidence-Driven Test-Time RL for Vision-Language-Action Models
T^2VLA is a test-time reinforcement learning framework for VLAs that uses internal confidence to define intrinsic rewards via similarity to high-confidence expert demonstrations and a dual-expert bootstrapping mechanism.
-
SSI-Policy: Learning Structured Scene Interfaces for Vision-Language Robotic Manipulation
SSI-Policy uses an RGB-only Structured Scene Interface to improve LIBERO benchmark performance by nearly 15% with only 10 demonstrations per task compared to prior methods.
-
EventVLA: Event-Driven Visual Evidence Memory for Long-Horizon Vision-Language-Action Policies
EventVLA introduces foundational visual anchors and a Keyframe Evidence Memory module that predicts future keyframe probabilities from VLA embeddings to improve long-horizon task success by an average of 40% on 17 simulation and 4 real-world tasks.
-
LARA: Latent Action Representation Alignment for Vision-Language-Action Models
LARA jointly optimizes LAM and VLA models via representation alignment to improve robotic manipulation performance using human videos.
-
AEGIS: A Backup Reflex for Physical AI
AEGIS uses activation probes for early-warning detection of high-risk steps in weak policies and selectively escalates to stronger policies, recovering 10.1% of lost trajectories on LIBERO-Spatial while activating the strong policy on only 38% of steps.
-
TempoVLA: Learning Speed-Controllable Vision-Language-Action Policies
TempoVLA learns a single VLA policy with controllable execution speed via variable-speed trajectory augmentation and explicit speed conditioning.
-
Flow-based Policy Adaptation without Policy Updates
GLOVES learns flow models from limited expert demonstrations to selectively correct actions from non-expert policies or operators toward expert distributions using reverse-flow OOD detection as an intervention gate.
-
AffordanceVLA: A Vision-Language-Action Model Empowering Action Generation through Affordance-Aware Understanding
AffordanceVLA proposes a VLA model with affordance-aware modules (Which2Act, Where2Act, How2Act) in a Mixture-of-Transformer trained in three stages to improve robotic manipulation.
-
Flash-WAM: Modality-Aware Distillation for World Action Models
Flash-WAM introduces modality-specific consistency parametrizations to distill joint video-action diffusion models to single-step inference, delivering 23x speedup with preserved benchmark performance.
-
CLAW: Learning Continuous Latent Action World Models via Adversarial Latent Regularization
CLAW is an end-to-end self-supervised method that learns semantically meaningful continuous latent actions and predictive world models from action-free videos to support imitation learning and goal-directed planning.
-
Revisiting Embodied Chain-of-Thought for Generalizable Robot Manipulation
ERVLA trains on a 978k-trajectory embodied CoT corpus using reasoning as supervision with dropout, then predicts actions without CoT at test time, reaching 86.9% on LIBERO-Plus and 53.2% on VLABench.