bugs about mae-pytorch

[launchauto]

MAE-pytorch/modeling_pretrain.py", line 296, in pretrain_mae_base_patch16_224
    **kwargs)
TypeError: __init__() got an unexpected keyword argument 'num_classes'

command
pip install --upgrade timm
will help

[pengzhiliang]

I'm sorry that I can't reproduce the bug you mentioned with the unmodified code in this repo. Can you provide more information?

[cuge1995]

just clone your repo, and follow the running code

# Set the path to save checkpoints
OUTPUT_DIR='output/'
# path to imagenet-1k train set
DATA_PATH='../ImageNet_ILSVRC2012/train'


OMP_NUM_THREADS=1 python -m torch.distributed.launch --nproc_per_node=8 run_mae_pretraining.py \
        --data_path ${DATA_PATH} \
        --mask_ratio 0.75 \
        --model pretrain_mae_base_patch16_224 \
        --batch_size 128 \
        --opt_betas 0.9 0.95 \
        --warmup_epochs 40 \
        --epochs 1600 \
        --output_dir ${OUTPUT_DIR}

[cuge1995]

more logs are as follows:

Creating model: pretrain_mae_base_patch16_224
Traceback (most recent call last):
  File "run_mae_pretraining.py", line 264, in <module>
    main(opts)
  File "run_mae_pretraining.py", line 150, in main
    model = get_model(args)
  File "run_mae_pretraining.py", line 129, in get_model
    drop_block_rate=None,
  File "timm/models/factory.py", line 57, in create_model
    model = create_fn(**model_args, **kwargs)
  File "MAE-pytorch/modeling_pretrain.py", line 296, in pretrain_mae_base_patch16_224
    **kwargs)
TypeError: __init__() got an unexpected keyword argument 'num_classes'
Traceback (most recent call last):

[LiewFeng]

I can run it successfully. Maybe you can check the version of timm.

[chenhao-zju]

MAE-pytorch/modeling_pretrain.py", line 296, in pretrain_mae_base_patch16_224
**kwargs)
TypeError: init() got an unexpected keyword argument 'num_classes'

Have you solved this problem? I met it too

[pengzhiliang]

Hello, @shandong-university-weihai-chen . You can have a try:

pip install timm==0.3.2

[chenhao-zju]

It’s very nice of you, I solve this problem by changing all ‘>=’ to ‘==’ in ‘requirements.txt’

[pengzhiliang]

You are welcome.