百科.dev
全部条目AI 编程趋势榜开源项目技术资讯提交条目
登录
< 返回工具列表
P

proxylessnas

> DevOps
开源

[ICLR 2019] ProxylessNAS: 在目标任务和硬件上进行无代理的神经架构搜索

1.4K stars0 点赞0 次浏览
访问官网GitHub

工具介绍

[ICLR 2019] ProxylessNAS: 在目标任务和硬件上进行无代理的神经架构搜索

# ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware [[arXiv]](https://arxiv.org/abs/1812.00332) [[Poster]](assets/ProxylessNAS_iclr_poster_final.pdf) ```bash @inproceedings{ cai2018proxylessnas, title={Proxyless{NAS}: Direct Neural Architecture Search on Target Task and Hardware}, author={Han Cai and Ligeng Zhu and Song Han}, booktitle={International Conference on Learning Representations}, year={2019}, url={https://arxiv.org/pdf/1812.00332.pdf}, } ``` ## News - ProxylessNAS is integrated into [PytorchHub](https://pytorch.org/hub/pytorch_vision_proxylessnas/). - ProxylessNAS is integrated into Microsoft [NNI](https://nni.readthedocs.io/en/v1.6/NAS/Proxylessnas.html). - ProxylessNAS is integrated into Amazon [AutoGluon](https://auto.gluon.ai/0.3.0/tutorials/nas/enas_proxylessnas.html). - First place in the Visual Wake Words Challenge, TF-lite track, @CVPR 2019 - Third place in the Low Power Image Recognition Challenge (LPIRC), classification track, @CVPR 2019 ## Performance Without any proxy, directly and efficiently search neural network architectures on your target **task** and **hardware**! Now, proxylessnas is on [PyTorch Hub](https://pytorch.org/hub/pytorch_vision_proxylessnas/). You can load it with only two lines! ```python target_platform = "proxyless_cpu" # proxyless_gpu, proxyless_mobile, proxyless_mobile14 are also avaliable. model = torch.hub.load('mit-han-lab/ProxylessNAS', target_platform, pretrained=True) ```

Mobile settings GPU settings
| Model | Top-1 | Top-5 | Latency | |----------------------|----------|----------|---------| | MobilenetV2 | 72.0 | 91.0 | 6.1ms | | ShufflenetV2(1.5) | 72.6 | - | 7.3ms | | ResNet-34 | 73.3 | 91.4 | 8.0ms | | MNasNet(our impl) | 74.0 | 91.8 | 6.1ms | | ProxylessNAS (GPU) | 75.1 | 92.5 | 5.1ms |
ProxylessNAS(Mobile) consistently outperforms MobileNetV2 under various latency settings. ProxylessNAS(GPU) is 3.1% better than MobilenetV2 with 20% faster.
## Specialization People used to deploy one model to all platforms, but this is not good. To fully exploit the efficiency, we should specialize architectures for each platform. We provide a [visualization](assets/visualization.mp4) of search process. Please refer to our [paper](https://arxiv.org/abs/1812.00332) for more results. # How to use / evaluate * Use ```python # pytorch from proxyless_nas import proxyless_cpu, proxyless_gpu, proxyless_mobile, proxyless_mobile_14, proxyless_cifar net = proxyless_cpu(pretrained=True) # Yes, we provide pre-trained models! ``` ```python # tensorflow from proxyless_nas_tensorflow import proxyless_cpu, proxyless_gpu, proxyless_mobile, proxyless_mobile_14 tf_net = proxyless_cpu(pretrained=True) ``` If the above scripts failed to download, you download it manually from [Google Drive](https://drive.google.com/drive/folders/1qIaDsT95dKgrgaJk-KOMu6v9NLROv2tz?usp=sharing) and put them under `$HOME/.torch/proxyless_nas/`. * Evaluate `python eval.py --path 'Your path to imagent' --arch proxyless_cpu # pytorch ImageNet` `python eval.py -d cifar10 # pytorch cifar10 ` `python eval_tf.py --path 'Your path to imagent' --arch proxyless_cpu # tensorflow` ## File structure * [search](./search): code for neural architecture search. * [training](./training): code for training searched models. * [proxyless_nas_tensorflow](./proxyless_nas_tensorflow): pretrained models for tensorflow. * [proxyless_nas](./proxyless_nas): pretrained models for PyTorch. ## Projects with ProxylessNAS: * [ProxylessGaze](./proxyless_gaze/): Real-time Gaze Estimation with ProxylessNAS ## Related work on automated model compression and acceleration: [Once for All: Train One Network and Specialize it for Efficient Deployment](https://arxiv.org/abs/1908.09791) (ICLR'20, [code](https://github.com/mit-han-lab/once-for-all)) [ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware](https://arxiv.org/pdf/1812.00332.pdf) (ICLR’19) [AMC: AutoML for Model Compression and Acceleration on Mobile Devices](https://arxiv.org/pdf/1802.03494.pdf) (ECCV’18) [HAQ: Hardware-Aware Automated Quantization](https://arxiv.org/pdf/1811.08886.pdf) (CVPR’19, oral) [Defenstive Quantization: When Efficiency Meets Robustness](https://openreview.net/pdf?id=ryetZ20ctX) (ICLR'19)

Issues· 2 开放

查看全部 Issues在 GitHub 打开

暂无开放 Issues,或尚未同步最近议题。

> 标签

C++accelerationautomlefficient-modelhardware-aware

暂无评论,来聊聊你的看法吧

> 工具信息

发布日期2026年8月1日
最后更新2026年9月17日
分类DevOps
定价开源

> 相关工具

D
Docker
容器化平台,标准化应用交付
G
GitHub Actions
GitHub 原生 CI/CD 工作流
N
Nginx
高性能 Web 服务器与反向代理