Performance comparison of deep learning frameworks in image classification problems using convolutional and recurrent networks

Fonnegra, Rubén D.; Blair, Bryan; Díaz, Gabriel

doi:10.1109/colcomcon.2017.8088219

Cited by 18 publications

(18 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The authors did not inform the used CNN architectures. Fonnegra et al (2017) evaluated and compared the following libraries: TensorFlow, Theano and Torch. The comparison is performed through the implementation of recurrent and convolutional architectures for classifying images of two datasets: MNIST and CIFAR-10.…”

Section: Related Workmentioning

confidence: 99%

Performance Analysis of Deep Learning Libraries: TensorFlow and PyTorch

Florencio¹,

Valenç²,

Moreno³

et al. 2019

Journal of Computer Science

View full text Add to dashboard Cite

Through the increase in deep learning study and use, in the last years there was a development of specific libraries for Deep Neural Network (DNN). Each one of these libraries has different performance results and applies different techniques to optimize the implementation of algorithms. Therefore, even though implementing the same algorithm and using different libraries, the performance of their executions may have a considerable variation. For this reason, developers and scientists that work with deep learning need scientific experimental studies that examine the performance of those libraries. Therefore, this paper has the aim of evaluating and comparing these two libraries: TensorFlow and PyTorch. We have used three parameters: Hardware utilization, hardware temperature and execution time in a context of heterogeneous platforms with CPU and GPU. We used the MNIST database for training and testing the LeNet Convolutional Neural Network (CNN). We performed a scientific experiment following the Goal Question Metrics (GQM) methodology, data were validated through statistical tests. After data analysis, we show that PyTorch library presented a better performance, even though the TensorFlow library presented a greater GPU utilization rate.

show abstract

Section: Related Workmentioning

confidence: 99%

Performance Analysis of Deep Learning Libraries: TensorFlow and PyTorch

Florencio¹,

Valenç²,

Moreno³

et al. 2019

Journal of Computer Science

View full text Add to dashboard Cite

show abstract

“…A brief overview of the most commonly used DL software frameworks for computer vision applications (especially relevant to crack detection) is presented in this section, highlighting their pros and cons. This is mostly based on some recently published comparative studies on DL software frameworks by evaluating these frameworks in terms of extensibility, hardware utilization, and speed [9][10][11]. It is worth noting that all these DL frameworks are undergoing constant development with active contributions from researchers and the open-source community, and therefore the study results and conclusions from the reported comparative studies may not be up to date.…”

Section: Some Existing and Emerging Deep Learning Framework For Compmentioning

confidence: 99%

“…TensorFlow, originally developed by researchers and engineers working on the Google Brain Team, uses data flow graphs for numerical computation and is mainly designed for developing and implementing deep neural network models [14]. One major advantage of TensorFlow that vastly increased its popularity among DL researchers and companies is its ability to deploy computation to one or more CPUs/GPUs on a variety of systems and devices through a single application programming interface (API) [10]. Based on a comparative study of Theano, Torch, Neon, and TensorFlow DL frameworks, Bahrampour et al [9] concluded that TensorFlow, although a very flexible framework, is not as competitive as other studied frameworks in terms of its performance on a single GPU.…”

Section: Tensorflowmentioning

confidence: 99%

“…However, in a similar benchmarking study comparing Caffe, CNTK, TensorFlow, and Torch, Shi et al [11] reported that no single framework consistently outperforms others. Fonnegra et al [10] reported that TensorFlow was faster than Theno when the DL architecture contained long short term memory (LSTM) units as its core.…”

Section: Tensorflowmentioning

confidence: 99%

“…Since its initial release in 2002, Torch has grown to be a popular open-source DL framework for commercial applications (Facebook, IBM, etc. ), academic research studies, and so on, owing to its simplicity and extensibility [10]. Based on the results of a comparative study by Bahrampour et al [9], Torch performed the best for CPU-based training and deployment of any DL architecture that was tested.…”

Section: Torchmentioning

confidence: 99%

See 2 more Smart Citations

Deep Learning in Data-Driven Pavement Image Analysis and Automated Distress Detection: A Review

Gopalakrishnan

2018

Data

132

View full text Add to dashboard Cite

Deep learning, more specifically deep convolutional neural networks, is fast becoming a popular choice for computer vision-based automated pavement distress detection. While pavement image analysis has been extensively researched over the past three decades or so, recent ground-breaking achievements of deep learning algorithms in the areas of machine translation, speech recognition, and computer vision has sparked interest in the application of deep learning to automated detection of distresses in pavement images. This paper provides a narrative review of recently published studies in this field, highlighting the current achievements and challenges. A comparison of the deep learning software frameworks, network architecture, hyper-parameters employed by each study, and crack detection performance is provided, which is expected to provide a good foundation for driving further research on this important topic in the context of smart pavement or asset management systems. The review concludes with potential avenues for future research; especially in the application of deep learning to not only detect, but also characterize the type, extent, and severity of distresses from 2D and 3D pavement images.

show abstract

An experimental comparison of the widely used pre‐trained deep neural networks for image classification tasks towards revealing the promise of transfer‐learning

Kabakuş

Erdoğmuş

2022

Concurrency and Computation

View full text Add to dashboard Cite

Summary The easiest way to propose a solution based on deep neural networks is using the pre‐trained models through the transfer‐learning technique. Deep learning platforms provide various pre‐trained deep neural networks that can be easily applied for image classification tasks. So, “Which pre‐trained model provides the best performance for image classification tasks?” is a question that instinctively comes to mind and should be shed light on by the research community. To this end, we propose an experimental comparison of the six popular pre‐trained deep neural networks, namely, (i) VGG19, (ii) ResNet50, (iii) DenseNet201, (iv) MobileNetV2, (v) InceptionV3, and (vi) Xception by employing them through the transfer‐learning technique. Then, the proposed benchmark models were both trained and evaluated under the same configurations on two gold‐standard datasets, namely, (i) CIFAR‐10 and (ii) Stanford Dogs to benchmark them. Three evaluation metrics were employed to measure performance differences between the employed pre‐trained models as follows: (i) Accuracy, (ii) training duration, and (iii) inference time. The key findings that were obtained through the conducted a wide variety of experiments were discussed.

show abstract

Performance comparison of deep learning frameworks in image classification problems using convolutional and recurrent networks

Cited by 18 publications

References 4 publications

Performance Analysis of Deep Learning Libraries: TensorFlow and PyTorch

Performance Analysis of Deep Learning Libraries: TensorFlow and PyTorch

Deep Learning in Data-Driven Pavement Image Analysis and Automated Distress Detection: A Review

An experimental comparison of the widely used pre‐trained deep neural networks for image classification tasks towards revealing the promise of transfer‐learning

Contact Info

Product

Resources

About