mmpretrain
mmpretrain copied to clipboard
open-mmlab
[Bug] I encountered a problem while fine-tuning the model on a custom dataset.
Branch
main branch (mmpretrain version)
Describe the bug
I followed the instructions at https://mmpretrain.readthedocs.io/zh_CN/latest/notes/finetune_custom_dataset.html to fine-tune the model, but I encountered an issue in the custom dataset: "init() got an unexpected keyword argument 'split'". How can I resolve this problem? I noticed that someone else has also faced the same issue. Can you tell me how to fix it? The link is here: "https://github.com/open-mmlab/mmpretrain/discussions/1696".
Environment
base = [ '../base/models/resnet50.py', '../base/datasets/imagenet_bs32.py', '../base/schedules/imagenet_bs256.py', '../base/default_runtime.py' ]
模型设置
model = dict( backbone=dict( frozen_stages=2, init_cfg=dict( type='Pretrained', checkpoint='https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_8xb32_in1k_20210831-ea4938fc.pth', prefix='backbone', )), head=dict(num_classes=12), )
数据设置
data_root = 'data/custom_dataset' train_dataloader = dict( batch_size=256, dataset=dict( type='CustomDataset', data_root=data_root, ann_file='', data_prefix='train', with_label=True, )) val_dataloader = dict( dataset=dict( type='CustomDataset', data_root=data_root, ann_file='', data_prefix='val', with_label=True, )) test_dataloader = dict( dataset=dict( type='CustomDataset', data_root=data_root, ann_file='', data_prefix='test', with_label=True, ))
test_dataloader = val_dataloader
训练策略设置
optim_wrapper = dict( optimizer=dict(type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0001)) param_scheduler = dict( type='MultiStepLR', by_epoch=True, milestones=[15], gamma=0.1)
Other information
No response
Is a fix for this gonna land? I'm just following the basic custom dataset docs that the OP mentioned. This is my first time using the library and it doesn't give me warm-fuzzy's that the tutorial has been broken since at least August
