lightweight-neural-architecture-search

lightweight-neural-architecture-search copied to clipboard

Open MPI is not supported under windows.

LT1st

Thanks for this amazing repo. I'm currently working on training an efficient low-precision backbone and deploying it on an ARM Cortex-M7 MCU device with limited resources (512kB RAM, 2MB Flash)....

erectbranch

Supported device？

3

As shown in the config file `config_nas.py`, is it only the V100 or t40 are supported here? If i want to use 2 * 2080Ti, how can i change that?

David-19940718

Thanks for sharing the code of such an interesting work! Just got a few questions when reproducing the code: 1. The [links](https://github.com/alibaba/lightweight-neural-architecture-search/blob/main/configs/classification/README.md?plain=1#L50) to the pre-trained image classification models (RXX-like.pth) seem...

zhuojinl

how much memory was used

1

in paper said that we can use cpu with small memory source ,but when i run `sh tools/dist_search.sh configs/classification/deepmad_29M_224.py` 32G memory was used in 10 sec and the job has...

YiChuangDai

请问可以用于图像融合吗

2

liushh39

the limit of btn

1

The constraint limit the btn in https://github.com/alibaba/lightweight-neural-architecture-search/blob/6bf4d6949ed690b8ef59bcb843e2d36d03ebecd1/tinynas/spaces/mutator/super_res_k1kx_mutator.py#L72 https://github.com/alibaba/lightweight-neural-architecture-search/blob/6bf4d6949ed690b8ef59bcb843e2d36d03ebecd1/tinynas/spaces/mutator/super_res_k1kx_mutator.py#L82 but this https://github.com/tinyvision/DAMO-YOLO/blob/master/damo/base_models/backbones/nas_backbones/tinynas_nano_middle.txt seems to have broken the limit May I ask why? thanks

ct-wei

For DIY use

1

If I want to use the Deepmap method to create a specific parameter quantity resnet or create my own resnet101, how should I call the script? Is this open source

TheShy-Dream

The results are different after two searches

1

Hello, thanks for your great work. I run scripts/damo-yolo/example_k1kx_small.sh twice, two best_structure.txt results are different, and is different from your damo-yolo-s structure。 I wonder if it is normal?

XA-hyy

hey, in code 'get_deepmad_forward' func, the deepmad compute way is log(sqrt(c*k^2/g)), [https://github.com/alibaba/lightweight-neural-architecture-search/blob/main/tinynas/models/blocks_cnn_2d/blocks_basic.py#L444] but in paper, the definition of the CNN entropy is without sqrt option, 

Yuxinyi-Qiyu