DetNAS ThunderNet

 

DetNAS是backbone，旷世的首个搜索物体检测Backbone的方法

https://github.com/megvii-model/DetNAS

 

https://github.com/megvii-model/DetNAS/blob/d2298d7b05b2ea145f373209a0500617b0336ea9/maskrcnn_benchmark/modeling/backbone/detnasnet.py

 

1070上640 batch：6 30ms

我把后面通道调少了

import torch
import torch.nn as nn
from shuffle_blocks import ConvBNReLU, ShuffleNetV2BlockSearched, blocks_key, FC
import numpy as np

class ShuffleNetV2DetNAS(nn.Module):
    def __init__(self, model_size='DETNAS-300M'):
        super(ShuffleNetV2DetNAS, self).__init__()
        print('Model size is {}.'.format(model_size))

        n_class = 1000
        if '300M' in model_size: