Michio JP

Michio JP

1629564829

HRFormer | High-Resolution Transformer for Dense Prediction

High-Resolution Transformer for Dense Prediction

This is the official implementation of High-Resolution Transformer (HRT). We present a High-Resolution Transformer (HRT) that learns high-resolution representations for dense prediction tasks, in contrast to the original Vision Transformer that produces low-resolution representations and has high memory and computational cost. We take advantage of the multi-resolution parallel design introduced in high-resolution convolutional networks (HRNet), along with local-window self-attention that performs self-attention over small non-overlapping image windows, for improving the memory and computation efficiency. In addition, we introduce a convolution into the FFN to exchange information across the disconnected image windows. We demonstrate the effectiveness of the High-Resolution Transformeron human pose estimation and semantic segmentation tasks.

  • The High-Resolution Transformer architecture:

An official implementation of the High-Resolution Transformer for Dense Prediction

Pose estimation

2d Human Pose Estimation

Results on COCO val2017 with detector having human AP of 56.4 on COCO val2017 dataset

BackboneInput SizeAPAP50AP75ARMARLARckptlogscript
HRT-S256x19274.0%90.2%81.2%70.4%80.7%79.4%ckptlogscript
HRT-S384x28875.6%90.3%82.2%71.6%82.5%80.7%ckptlogscript
HRT-B256x19275.6%90.8%82.8%71.7%82.6%80.8%ckptlogscript
HRT-B384x28877.2%91.0%83.6%73.2%84.2%82.0%ckptlogscript

Results on COCO test-dev with detector having human AP of 56.4 on COCO val2017 dataset

BackboneInput SizeAPAP50AP75ARMARLARckptlogscript
HRT-S384x28874.5%92.3%82.1%70.7%80.6%79.8%ckptlogscript
HRT-B384x28876.2%92.7%83.8%72.5%82.3%81.2%ckptlogscript

The models are first pre-trained on ImageNet-1K dataset, and then fine-tuned on COCO val2017 dataset.

Semantic segmentation

Cityscapes

Performance on the Cityscapes dataset. The models are trained and tested with input size of 512x1024 and 1024x2048 respectively.

MethodsBackboneWindow SizeTrain SetTest SetIterationsBatch SizeOHEMmIoUmIoU (Multi-Scale)Logckptscript
OCRNetHRT-S7x7TrainVal800008Yes80.081.0logckptscript
OCRNetHRT-B7x7TrainVal800008Yes81.482.0logckptscript
OCRNetHRT-B15x15TrainVal800008Yes81.982.6logckptscript

PASCAL-Context

The models are trained with the input size of 520x520, and tested with original size.

MethodsBackboneWindow SizeTrain SetTest SetIterationsBatch SizeOHEMmIoUmIoU (Multi-Scale)Logckptscript
OCRNetHRT-S7x7TrainVal6000016Yes53.854.6logckptscript
OCRNetHRT-B7x7TrainVal6000016Yes56.357.1logckptscript
OCRNetHRT-B15x15TrainVal6000016Yes57.658.5logckptscript

COCO-Stuff

The models are trained with input size of 520x520, and tested with original size.

MethodsBackboneWindow SizeTrain SetTest SetIterationsBatch SizeOHEMmIoUmIoU (Multi-Scale)Logckptscript
OCRNetHRT-S7x7TrainVal6000016Yes37.938.9logckptscript
OCRNetHRT-B7x7TrainVal6000016Yes41.642.5logckptscript
OCRNetHRT-B15x15TrainVal6000016Yes42.443.3logckptscript

ADE20K

The models are trained with input size of 520x520, and tested with original size. The results with window size 15x15 will be updated latter.

MethodsBackboneWindow SizeTrain SetTest SetIterationsBatch SizeOHEMmIoUmIoU (Multi-Scale)Logckptscript
OCRNetHRT-S7x7TrainVal1500008Yes44.045.1logckptscript
OCRNetHRT-B7x7TrainVal1500008Yes46.347.6logckptscript
OCRNetHRT-B13x13TrainVal1500008Yes48.750.0logckptscript
OCRNetHRT-B15x15TrainVal1500008Yes-----

Classification

Results on ImageNet-1K

Backboneacc@1acc@5#paramsFLOPsckptlogscript
HRT-T78.6%94.2%8.0M1.83Gckptlogscript
HRT-S81.2%95.6%13.5M3.56Gckptlogscript
HRT-B82.8%96.3%50.3M13.71Gckptlogscript

Citation

If you find this project useful in your research, please consider cite:

@article{YuanFHZCW21,
  title={HRT: High-Resolution Transformer for Dense Prediction},
  author={Yuhui Yuan and Rao Fu and Lang Huang and Chao Zhang and Xilin Chen and Jingdong Wang},
  booktitle={arXiv},
  year={2021}
}

Acknowledgment

This project is developed based on the Swin-Transformer, openseg.pytorch, and mmpose.

git diff-index HEAD
git subtree add -P pose  

Download Details:

Author: HRNet

Source Code: https://github.com/HRNet/HRFormer

 

What is GEEK

Buddha Community

Ajay Kapoor

1624252974

Digital Transformation Consulting Services & solutions

Compete in this Digital-First world with PixelCrayons’ advanced level digital transformation consulting services. With 16+ years of domain expertise, we have transformed thousands of companies digitally. Our insight-led, unique, and mindful thinking process helps organizations realize Digital Capital from business outcomes.

Let our expert digital transformation consultants partner with you in order to solve even complex business problems at speed and at scale.

Digital transformation company in india

#digital transformation agency #top digital transformation companies in india #digital transformation companies in india #digital transformation services india #digital transformation consulting firms

Ian  Robinson

Ian Robinson

1623223443

Predictive Modeling in Data Science

Predictive modeling is an integral tool used in the data science world — learn the five primary predictive models and how to use them properly.

Predictive modeling in data science is used to answer the question “What is going to happen in the future, based on known past behaviors?” Modeling is an essential part of data science, and it is mainly divided into predictive and preventive modeling. Predictive modeling, also known as predictive analytics, is the process of using data and statistical algorithms to predict outcomes with data models. Anything from sports outcomes, television ratings to technological advances, and corporate economies can be predicted using these models.

Top 5 Predictive Models

  1. Classification Model: It is the simplest of all predictive analytics models. It puts data in categories based on its historical data. Classification models are best to answer “yes or no” types of questions.
  2. Clustering Model: This model groups data points into separate groups, based on similar behavior.
  3. **Forecast Model: **One of the most widely used predictive analytics models. It deals with metric value prediction, and this model can be applied wherever historical numerical data is available.
  4. Outliers Model: This model, as the name suggests, is oriented around exceptional data entries within a dataset. It can identify exceptional figures either by themselves or in concurrence with other numbers and categories.
  5. Time Series Model: This predictive model consists of a series of data points captured, using time as the input limit. It uses the data from previous years to develop a numerical metric and predicts the next three to six weeks of data using that metric.

#big data #data science #predictive analytics #predictive analysis #predictive modeling #predictive models

Michio JP

Michio JP

1629564829

HRFormer | High-Resolution Transformer for Dense Prediction

High-Resolution Transformer for Dense Prediction

This is the official implementation of High-Resolution Transformer (HRT). We present a High-Resolution Transformer (HRT) that learns high-resolution representations for dense prediction tasks, in contrast to the original Vision Transformer that produces low-resolution representations and has high memory and computational cost. We take advantage of the multi-resolution parallel design introduced in high-resolution convolutional networks (HRNet), along with local-window self-attention that performs self-attention over small non-overlapping image windows, for improving the memory and computation efficiency. In addition, we introduce a convolution into the FFN to exchange information across the disconnected image windows. We demonstrate the effectiveness of the High-Resolution Transformeron human pose estimation and semantic segmentation tasks.

  • The High-Resolution Transformer architecture:

An official implementation of the High-Resolution Transformer for Dense Prediction

Pose estimation

2d Human Pose Estimation

Results on COCO val2017 with detector having human AP of 56.4 on COCO val2017 dataset

BackboneInput SizeAPAP50AP75ARMARLARckptlogscript
HRT-S256x19274.0%90.2%81.2%70.4%80.7%79.4%ckptlogscript
HRT-S384x28875.6%90.3%82.2%71.6%82.5%80.7%ckptlogscript
HRT-B256x19275.6%90.8%82.8%71.7%82.6%80.8%ckptlogscript
HRT-B384x28877.2%91.0%83.6%73.2%84.2%82.0%ckptlogscript

Results on COCO test-dev with detector having human AP of 56.4 on COCO val2017 dataset

BackboneInput SizeAPAP50AP75ARMARLARckptlogscript
HRT-S384x28874.5%92.3%82.1%70.7%80.6%79.8%ckptlogscript
HRT-B384x28876.2%92.7%83.8%72.5%82.3%81.2%ckptlogscript

The models are first pre-trained on ImageNet-1K dataset, and then fine-tuned on COCO val2017 dataset.

Semantic segmentation

Cityscapes

Performance on the Cityscapes dataset. The models are trained and tested with input size of 512x1024 and 1024x2048 respectively.

MethodsBackboneWindow SizeTrain SetTest SetIterationsBatch SizeOHEMmIoUmIoU (Multi-Scale)Logckptscript
OCRNetHRT-S7x7TrainVal800008Yes80.081.0logckptscript
OCRNetHRT-B7x7TrainVal800008Yes81.482.0logckptscript
OCRNetHRT-B15x15TrainVal800008Yes81.982.6logckptscript

PASCAL-Context

The models are trained with the input size of 520x520, and tested with original size.

MethodsBackboneWindow SizeTrain SetTest SetIterationsBatch SizeOHEMmIoUmIoU (Multi-Scale)Logckptscript
OCRNetHRT-S7x7TrainVal6000016Yes53.854.6logckptscript
OCRNetHRT-B7x7TrainVal6000016Yes56.357.1logckptscript
OCRNetHRT-B15x15TrainVal6000016Yes57.658.5logckptscript

COCO-Stuff

The models are trained with input size of 520x520, and tested with original size.

MethodsBackboneWindow SizeTrain SetTest SetIterationsBatch SizeOHEMmIoUmIoU (Multi-Scale)Logckptscript
OCRNetHRT-S7x7TrainVal6000016Yes37.938.9logckptscript
OCRNetHRT-B7x7TrainVal6000016Yes41.642.5logckptscript
OCRNetHRT-B15x15TrainVal6000016Yes42.443.3logckptscript

ADE20K

The models are trained with input size of 520x520, and tested with original size. The results with window size 15x15 will be updated latter.

MethodsBackboneWindow SizeTrain SetTest SetIterationsBatch SizeOHEMmIoUmIoU (Multi-Scale)Logckptscript
OCRNetHRT-S7x7TrainVal1500008Yes44.045.1logckptscript
OCRNetHRT-B7x7TrainVal1500008Yes46.347.6logckptscript
OCRNetHRT-B13x13TrainVal1500008Yes48.750.0logckptscript
OCRNetHRT-B15x15TrainVal1500008Yes-----

Classification

Results on ImageNet-1K

Backboneacc@1acc@5#paramsFLOPsckptlogscript
HRT-T78.6%94.2%8.0M1.83Gckptlogscript
HRT-S81.2%95.6%13.5M3.56Gckptlogscript
HRT-B82.8%96.3%50.3M13.71Gckptlogscript

Citation

If you find this project useful in your research, please consider cite:

@article{YuanFHZCW21,
  title={HRT: High-Resolution Transformer for Dense Prediction},
  author={Yuhui Yuan and Rao Fu and Lang Huang and Chao Zhang and Xilin Chen and Jingdong Wang},
  booktitle={arXiv},
  year={2021}
}

Acknowledgment

This project is developed based on the Swin-Transformer, openseg.pytorch, and mmpose.

git diff-index HEAD
git subtree add -P pose  

Download Details:

Author: HRNet

Source Code: https://github.com/HRNet/HRFormer

 

Chelsie  Towne

Chelsie Towne

1596716340

A Deep Dive Into the Transformer Architecture – The Transformer Models

Transformers for Natural Language Processing

It may seem like a long time since the world of natural language processing (NLP) was transformed by the seminal “Attention is All You Need” paper by Vaswani et al., but in fact that was less than 3 years ago. The relative recency of the introduction of transformer architectures and the ubiquity with which they have upended language tasks speaks to the rapid rate of progress in machine learning and artificial intelligence. There’s no better time than now to gain a deep understanding of the inner workings of transformer architectures, especially with transformer models making big inroads into diverse new applications like predicting chemical reactions and reinforcement learning.

Whether you’re an old hand or you’re only paying attention to transformer style architecture for the first time, this article should offer something for you. First, we’ll dive deep into the fundamental concepts used to build the original 2017 Transformer. Then we’ll touch on some of the developments implemented in subsequent transformer models. Where appropriate we’ll point out some limitations and how modern models inheriting ideas from the original Transformer are trying to overcome various shortcomings or improve performance.

What Do Transformers Do?

Transformers are the current state-of-the-art type of model for dealing with sequences. Perhaps the most prominent application of these models is in text processing tasks, and the most prominent of these is machine translation. In fact, transformers and their conceptual progeny have infiltrated just about every benchmark leaderboard in natural language processing (NLP), from question answering to grammar correction. In many ways transformer architectures are undergoing a surge in development similar to what we saw with convolutional neural networks following the 2012 ImageNet competition, for better and for worse.

#natural language processing #ai artificial intelligence #transformers #transformer architecture #transformer models

Otho  Hagenes

Otho Hagenes

1617419868

Top Five Artificial Intelligence Predictions For 2021

As AI becomes more ubiquitous, it’s also become more autonomous — able to act on its own without human supervision. This demonstrates progress, but it also introduces concerns around control over AI. The AI Arms Race has driven organizations everywhere to deliver the most sophisticated algorithms around, but this can come at a price, ignoring cultural and ethical values that are critical to responsible AI. Here are five predictions on what we should expect to see in AI in 2021:

  1. Something’s going to give around AI governance
  2. Most consumers will continue to be sceptical of AI
  3. Digital transformation (DX) finds its moment
  4. Organizations will increasingly push AI to the edge
  5. ModelOps will become the “go-to” approach for AI deployment.

#opinions #2021 ai predictions #ai predictions for 2021 #artificial intelligence predictions #five artificial intelligence predictions for 2021