baseline
tu12306 opened this issue · 4 comments
您好，仓库里没有baseline训练文件，请问您是怎么对照剪枝效果呢？baseline是这样子训练吗？
base = [
'../base/models/resnet50.py', '../base/datasets/cifar10_bs16.py',
'../base/schedules/cifar10_bs128.py', '../base/default_runtime.py'
]
optimizer = dict(lr=0.004)
work_dir = "work_dirs/resnet50-baselilne"
load_from = "https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_8xb32_in1k_20210831-ea4938fc.pth"
好的，谢谢。像我这种用cifar10数据集做实验，原有resnet50_prunig中 load_from的权重文件pth是不是不合适了呢？
对的，原来的ckpt是imagenet训的，分类head的维度和cifar不一样。你可以按照 https://github.com/open-mmlab/mmclassification/blob/master/configs/resnet/resnet50_8xb16_cifar10.py 修改一下config，load_from 改成 https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_b16x8_cifar10_20210528-f54bfad9.pth
好的，谢谢~