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%).
201
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
A prompt-only attack called command-preserving trajectory redirection can steer VLA robot behavior to attacker-chosen physical outcomes while the text still appears to match the intended task.
PROBEACT is a plug-and-play intervention framework that combines hidden-state probing, kinematic failure detection, and CBF-based correction to boost success rates of pre-trained VLA models on the LIBERO-plus benchmark from 69.6% to 74.1%.
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.
citing papers explorer
-
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.
-
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.
-
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.
-
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.
-
VLAFlow: A Unified Training Framework for Vision-Language-Action Models via Co-training and Future Latent Alignment
VLAFlow shows that combining language-supervised co-training with future latent alignment produces the most stable transfer performance for vision-language-action models trained on mixed robot data.
-
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.
-
FiberTune: Preserving Action-Fiber Visual Residuals in Vision-Language-Action Fine-Tuning
FiberTune is a new fine-tuning objective that preserves action-fiber visual residuals in VLA policies, yielding performance gains on simulation and physical robot 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.
-
Mitigating State Aliasing in Vision-Language-Action Models via Inverse Dynamics Learning
Inverse dynamics prediction is added as an auxiliary task to reduce state aliasing in VLA models by directly supervising the vision encoder on action-relevant visual distinctions using only standard observation-action pairs.
-
LAST: Bridging Vision-Language and Action Manifolds via Gromov-Wasserstein Alignment
LAST linearizes action manifolds with Lie-algebraic mapping and discretizes them into approximately isotropic charts to align with VL semantic geometry via Gromov-Wasserstein distance.
-
ActQuant: Sub-4-bit Action-Guided Quantization for Vision-Language-Action Models
ActQuant achieves sub-4-bit (down to 2.5 bpw) quantization of VLA models via action-contribution bit allocation and curvature-based scale tuning, retaining over 90% performance on LIBERO and physical robot tasks.
-
ForgeVLA: Federated Vision-Language-Action Learning without Language Annotations
ForgeVLA enables federated VLA model training from unlabeled vision-action pairs by recovering language via embodied classifiers and using contrastive planning plus adaptive aggregation to avoid feature collapse.
-
TriRelVLA: Triadic Relational Structure for Generalizable Embodied Manipulation
TriRelVLA introduces triadic object-hand-task relational representations and a task-grounded graph transformer with a relational bottleneck to improve generalization in robotic manipulation across scenes, objects, and tasks.
-
Test-Time Perturbation Learning with Delayed Feedback for Vision-Language-Action Models
PDF improves VLA success rates on LIBERO and Atari by applying test-time perturbation learning with delayed feedback to correct trajectory overfitting and overconfidence.
-
RoboStereo: Dual-Tower 4D Embodied World Models for Unified Policy Optimization
A dual-tower 4D embodied world model called RoboStereo reduces geometric hallucinations and delivers over 97% relative improvement on manipulation tasks via test-time augmentation, imitative learning, and open exploration.
-
VLANeXt: Recipes for Building Strong VLA Models
VLANeXt distills 12 design insights from a unified VLA study into a model that outperforms prior methods on LIBERO benchmarks while releasing code for further exploration.
-
Scaling by Diversified Experience for Vision-Language-Action Models
SyVLA uses Intention Decoupling and similar-sample guided RL on diversified experiences to improve VLA model task success and out-of-distribution generalization while keeping vision-language abilities.
-
Light-WAM: Efficient World Action Models with State-Fusion Action Decoding
Light-WAM is a lightweight world action model that performs latent-space video supervision and state-fusion action decoding to achieve usable multi-task robot performance with 0.44B parameters and low inference latency.
-
SAFE-Pruner: Semantic Attention-Guided Future-Aware Token Pruning for Efficient Vision-Language-Action Manipulation
SAFE-Pruner forecasts deep-layer token saliency in VLA models via semantic attention consistency and adaptive subtask detection to achieve up to 1.89x speedup with under 1.7% success rate loss.
-
Rethinking VLM Representation for VLA Initialization
Experiments indicate original VLM representations are crucial for VLA performance, LoRA outperforms full finetuning, and staged robot-data pretraining yields the strongest initialization.
-
QuoVLA: Quotient Space for Vision-Language-Action Models
QuoVLA introduces a quotient-space framework that compresses VLM latents into action-sufficient representations via quantization and dual-branch design for better VLA generalization.
-
Test-Time Training for Visual Foresight Vision-Language-Action Models
T³VF applies test-time training on natural future-prediction supervision pairs with adaptive filtering to mitigate OOD shifts in VF-VLA models at modest extra inference cost.
-
LiteVLA-H: Dual-Rate Vision-Language-Action Inference for Onboard Aerial Guidance and Semantic Perception
LiteVLA-H delivers 19.74 Hz action tokens and 6 Hz semantic outputs on Jetson Orin via dual-rate scheduling and mixed fine-tuning, outperforming recent VLA baselines in edge action rate while preserving descriptive competence.
-
SpanVLA: Efficient Action Bridging and Learning from Negative-Recovery Samples for Vision-Language-Action Model
SpanVLA reduces action generation latency via flow-matching conditioned on history and improves robustness by training on negative-recovery samples with GRPO and a dedicated reasoning dataset.
-
R3D: Revisiting 3D Policy Learning
A transformer 3D encoder plus diffusion decoder architecture, with 3D-specific augmentations, outperforms prior 3D policy methods on manipulation benchmarks by improving training stability.
-
HiVLA: A Visual-Grounded-Centric Hierarchical Embodied Manipulation System
HiVLA decouples VLM-based semantic planning with visual grounding from a cascaded cross-attention DiT action expert, outperforming end-to-end VLAs on long-horizon and fine-grained manipulation.
-
Causal World Modeling for Robot Control
LingBot-VA combines video world modeling with policy learning via Mixture-of-Transformers, closed-loop rollouts, and asynchronous inference to improve robot manipulation in simulation and real settings.
-
General Covariant Action Modeling: Constructing Generalized Manifolds via Spatio-Temporal Decoupling
GAM framework uses arc-length parameterization for temporal invariance and schema-affine factorization for geometric invariance to build a covariant action manifold integrated into VLA models for improved generalization from sparse data.
-
From Abstraction to Instantiation: Learning Behavioral Representation for Vision-Language-Action Model
BehaviorVLA learns long-horizon behavioral representations via causal Mamba encoder and phase-conditioned decoder, reporting SOTA results of 58% on RoboTwin 2.0, 98% on LIBERO, 4.36 on CALVIN, and matching OpenVLA-OFT performance with 50% data in sim-to-real transfer.
-
Evo-Depth: A Lightweight Depth-Enhanced Vision-Language-Action Model
Evo-Depth is a compact VLA model using a lightweight implicit depth encoder from RGB views plus progressive alignment to boost manipulation performance without added hardware.
- Towards Generalizable Robotic Manipulation in Dynamic Environments