2408.03326] LLaVA-OneVision: Easy Visual Task Transfer Skip to main content We gratefully acknowledge support

2408.03326] LLaVA-OneVision: Easy Visual Task Transfer Skip to main content We gratefully acknowledge support

00369] How to Train Your Fact Verifier: Knowledge Transfer with Multimodal Open Models Skip to main content

00369] How to Train Your Fact Verifier: Knowledge Transfer with Multimodal Open Models Skip to main content

2502.10377] ReStyle3D: Scene-Level Appearance Transfer with Semantic Correspondences arXiv Is Hiring a DevOps

2502.10377] ReStyle3D: Scene-Level Appearance Transfer with Semantic Correspondences Happy Open Access Week from arXiv! YOU

the Pre-train Basin: Insights on Ensembling in Transfer Learning Skip to main content We gratefully acknowledge

Multi-View Latent Optimization for Pose-Invariant Hairstyle Transfer Happy Open Access Week from arXiv! YOU make open access

2108.12847] Non-Parametric Neural Style Transfer Skip to main content We gratefully acknowledge support

2205.02841] Understanding Transfer Learning for Chest Radiograph Clinical Report Generation with

2303.09665] LOCATE: Localize and Transfer Object Parts for Weakly Supervised Affordance Grounding Skip

2303.09665] LOCATE: Localize and Transfer Object Parts for Weakly Supervised Affordance Grounding Skip

Generalization for Image Recognition via Cross-Client Style Transfer Skip to main content We gratefully acknowledge support

09838] BioSpark: Beyond Analogical Inspiration to LLM-augmented Transfer Skip to main content We gratefully acknowledge support

09838] BioSpark: Beyond Analogical Inspiration to LLM-augmented Transfer Skip to main content We gratefully acknowledge support

cond-mat/0702502] Spatial Transportation Networks with Transfer Costs: Asymptotic Optimality of Hub and Spoke

2210.13702] DeXtreme: Transfer of Agile In-hand Manipulation from

Tuning for Parameter and Memory Efficient Transfer Learning Skip to main content We gratefully acknowledge

2112.06825] VL-Adapter: Parameter-Efficient Transfer Learning for Vision-and-Language Tasks

1710.00756] Progressive Color Transfer with Dense Semantic Correspondences Skip to main content

1910.10683] Exploring the Limits of Transfer Learning with a Unified Text-to-Text

task misses an opportunity for cross-task knowledge transfer. We propose Improvise to Initialize (I2I), a

a novel image-guided texture synthesis method to transfer relightable textures from a small number of

a novel image-guided texture synthesis method to transfer relightable textures from a small number of

outstanding performance in many types of transfer learning: fully sampled fine-tuning, linear probing, few-shot transfer

outstanding performance in many types of transfer learning: fully sampled fine-tuning, linear probing, few-shot transfer

shared neural network parameters, where efficiency may be improved through transfer across related tasks. In practice, however, this is

shared neural network parameters, where efficiency may be improved through transfer across related tasks. In practice, however, this is

to enable zero-shot and few-shot transfer of the pre-trained models. We study

to enable zero-shot and few-shot transfer of the pre-trained models. We study

competitively to state-of-the-art transfer-learning methods requiring greater levels of supervision. We

of the learned dense correspondences through segmentation mask transfer on multiple datasets. We also show that the proposed

multimodal task learning, but do not enable cross-task knowledge transfer. We envision that CLiMB will facilitate research on a

to a notion of generative texture transfer. Comments: CVPR 2024. Code and additional visualizations available

specific features to be encapsulated, while still enabling transfer to multiple cities. We present an interactive navigation

specific features to be encapsulated, while still enabling transfer to multiple cities. We present an interactive navigation

classification in both zero-shot and transfer learning settings. However, we show that directly applying such models

classification in both zero-shot and transfer learning settings. However, we show that directly applying such models

is specialized in one domain, to transfer detailed domain-specific knowledge into the student universal

estimation. Such structural and contact priors can easily transfer to real-world data with barely any domain

training and evaluating simulation-to-real transfer for a novel vision task: unary and

to capture identity-disentangled facial expressions, facilitating accurate expression transfer for enhanced realism in animated scenes

suitability of the dataset for transfer learning by showing that image embeddings trained on it achieve

is specialized in one domain, to transfer detailed domain-specific knowledge into the student universal

to capture identity-disentangled facial expressions, facilitating accurate expression transfer for enhanced realism in animated scenes

suitability of the dataset for transfer learning by showing that image embeddings trained on it achieve

and trained to be promptable, so it can transfer zero-shot to new image distributions and

our new system shows significant superiority, especially in transfer learning. Our code and data will be made

multitask policies pretrained on GPT4-generated simulation tasks exhibit stronger transfer to unseen long-horizon tasks in the

and trained to be promptable, so it can transfer zero-shot to new image distributions and

and employ 3D-2D alignment techniques to transfer materials to different parts of each

deals with the problem of knowledge transfer between two datasets with domain-shift as well as category

data in object detection training and transfer knowledge between domains via pseudo labels. We further propose the

to $95\%$ fewer weights. Furthermore, we devise a transfer learning setting for ENEs, by permuting all features

in pre-training can benefit in-domain (ID) transfer performance, where the training and testing

processes of gene birth (duplication and transfer) and death (loss) to explain the

and exhibits strong sim-to-real zero-shot transfer. Combining the proposed pipeline and the

and exhibits strong sim-to-real zero-shot transfer. Combining the proposed pipeline and the

robotics, out-of-distribution (OOD) generalization, continual learning, transfer learning, uncertainty estimation / calibration, OOD detection, adversarial robustness, distillation, sparsity

in a common feature space, which allows zero-shot transfer to a downstream task via prompting, i.e.

generation, and even edit prompts for style transfer! Finally, being a gradient-free optimization approach, MILS

limited to specific editing types (e.g., object overlay, style transfer), or apply to synthetically generated images, or require

studied and benchmarked, it remains challenging to transfer this success to other skeleton structures with limited

generation, and even edit prompts for style transfer! Finally, being a gradient-free optimization approach, MILS

this end, we devise a cross-view knowledge transfer framework that extracts affordance-specific features from exocentric interactions and

studied and benchmarked, it remains challenging to transfer this success to other skeleton structures with limited

end, we devise a cross-view affordance knowledge transfer framework that extracts affordance-specific features from exocentric interactions and

estimation. Such structural and contact priors can easily transfer to real-world data with barely any domain

other authors View PDF HTML (experimental) Abstract: We introduce Cosmos-Transfer, a conditional world generation model that can generate

the models more knowledge efficient such as multi-task learning, transfer learning, weakly supervised and unsupervised learning etc. This

including mesh-to-video generation, camera control, motion transfer, and object manipulation. Comments: Project page: this https

evaluate models effectively on vision-centric tasks such as style transfer, a challenging area for current models

to fool trained models. Adversarial examples often exhibit black-box transfer, meaning that adversarial examples for one model can

on this data, which we call RT-X, exhibits positive transfer and improves the capabilities of

models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities

FeatureNeRF on tasks of 2D/3D semantic keypoint transfer and 2D/3D object part segmentation. Our extensive

Abstract: Reinforcement learning combined with sim-to-real transfer offers a general framework for developing

in simulation using these refined motions and transfer it to the real humanoid robot in

approach in terms of value function transfer, learning within redundant action spaces, and learning off

including mesh-to-video generation, camera control, motion transfer, and object manipulation. Comments: Project page: this https

Abstract: Prompt tuning is a new few-shot transfer learning technique that only tunes the learnable prompt

the use of power beaming to transfer energy and accelerate spacecraft. Applications suggested for

face geometry. In the second texture style transfer stage, we focus on performing style transfer on the canonical texture by adopting afirst framework for one-shot 3D portrait style transfer, which can generate 3D face models with both the

sets and performs the only possible legal transfer of a disk between the chosen

solve complex sequences of tasks--while both leveraging transfer and avoiding catastrophic forgetting--remains a key

solve complex sequences of tasks--while both leveraging transfer and avoiding catastrophic forgetting--remains a key

enforces a time-lock on the transfer of control of funds to

in machine learning, such as generative modeling and transfer learning. In this survey we explore contributions of

policies in simulation and perform Sim2Real transfer for real robot deployment. We perform extensive experiments

learned visual concepts (or describe new ones) enabling zero-shot transfer of the model to downstream

modalities, on multiple embodiments, and further, exhibits positive transfer: the model benefits from diverse joint training across

effectively with real-world dynamics. We evaluate ASAP across three transfer scenarios: IsaacGym to IsaacSim, IsaacGym to Genesis

images across 100k unique identities. For precise appearance transfer, we introduce a transformer encoder-decoder architecture adapted

Abstract: Reinforcement learning combined with sim-to-real transfer offers a general framework for developing