WXinlong / SOLO

SOLO and SOLOv2 for instance segmentation, ECCV 2020 & NeurIPS 2020.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

How to test the calculation amount (Flops) of solov2? Thank you~

solauky opened this issue · comments

Hello, how should I test the calculation amount (Flops) of solov2 in mmdetection2.5? How to debug? I used "get_flops.py" but the result showed "NotImplementedError: forward_dummy is not implemented in SOLOv2". Thank you~

Traceback (most recent call last):
File "tools/analysis_tools/get_flops.py", line 97, in
main()
File "tools/analysis_tools/get_flops.py", line 82, in main
flops, params = get_model_complexity_info(model, input_shape)
File "/home/scsc01/anaconda3/envs/lxx220/lib/python3.7/site-packages/mmcv/cnn/utils/flops_counter.py", line 105, in get_model_complexity_info
_ = flops_model(batch)
File "/home/scsc01/anaconda3/envs/lxx220/lib/python3.7/site-packages/torch/nn/modules/module.py", line 1071, in _call_impl
result = forward_call(*input, **kwargs)
File "/home/scsc01/lxx220/mmdet/models/detectors/single_stage_instance_seg.py", line 69, in forward_dummy
f'forward_dummy is not implemented in {self.class.name}')
NotImplementedError: forward_dummy is not implemented in SOLOv2