Torchvision the machine-vision package of torch

By Sébastien Marcel, Yann Rodriguez. 2010

603 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Mitigating Sybils in Federated Learning Poisoning (2018, methodology): “For VGGFace2, we use two popular architectures pre-trained on Imagenet from the torchvision package [25]: SqueezeNet1.”
NSML: Meet the MLaaS platform with a real-world case study (2018, methodology): “All the programming codes are written in python using PyTorch (Paszke et al., 2017) and Torchvision (Marcel & Rodriguez, 2010) libraries.”
Partial success in closing the gap between human and machine vision (2021, methodology): “In order to disentangle the influence of objective function, architecture and training dataset size, we tested a total of 52 models: 24 standard ImageNet-trained CNNs [37], 8 selfsupervised models [38–43],2 6 Big Transfer models [45], 5 adversarially trained models [46], 5 vision transformers [47, 4”
Engaging Image Chat: Modeling Personality in Grounded Dialogue (2018, methodology): “We used the implementation provided in the torchvision project (Marcel and Rodriguez, 2010).”
Federated Multi-Task Learning under a Mixture of Distributions (2021, methodology): “We also used Torchvision [40] implementation of MobileNet-v2 [50], and for image datasets preprossessing.”
Enhancing Adversarial Example Transferability With an Intermediate Level Attack (2019, methodology): “We also tested ILAP on ImageNet, with ResNet18, DenseNet121, SqueezeNet, and AlexNet5 pretrained on ImageNet (as provided in [17]).”
Engaging Image Captioning via Personality (2018, methodology): “We used the implementation provided in the torchvision project [33].”
Does Knowledge Distillation Really Work? (2021, methodology): “Conventional Image Transformations Standard data augmentations used in computer vision [30]: random rotations by up to 20 degrees, random vertical flips, color jitter and all possible combinations.”
Evaluation of Neural Architectures Trained with Square Loss vs Cross-Entropy in Classification Tasks (2020, methodology): “It is the default for classification in widely used packages such as NLP implementation HuggingFace’s Transformers [WDS19], speech classification by ESPnet [WHK18] and image classification implemented by torchvision [MR10].”
Throughput-Optimal Topology Design for Cross-Silo Federated Learning (2020, methodology): “In particular we used the torchvision [70] implementation of ResNet-18.”
A Simple Way to Make Neural Networks Robust Against Diverse Image Corruptions (2020, methodology): “We use various open source software packages for our experiments, most notably Docker [24], scipy and numpy [41], PyTorch [31] and torchvision [23].”
Distribution-Free, Risk-Controlling Prediction Sets (2021, methodology): “We use a pretrained ResNet-152 from the torchvision repository as the base model π̂x [43, 44].”
PopSGD: Decentralized Stochastic Gradient Descent in the Population Model (2019, methodology): “[34] Sébastien Marcel and Yann Rodriguez.”
Toward Transformer-Based Object Detection (2020, methodology): “Unless otherwise noted, we use the torchvision [26]”
Image-Chat: Engaging Grounded Conversations (2018, methodology): “We used the implementation provided in the torchvision project (Marcel and Rodriguez, 2010).”