Mmclassification installation and debugging

        MMClassification It's based on PyTorch Open source image classification toolbox , yes OpenMMLab One of the members of the project The main branch code currently supports PyTorch 1.5 Version above . MMClassification It contains many classic or high-performance image classification algorithms , Some of the latest papers and achievements are reproduced , In particular, it includes a large number of CVPR Reproduction of papers .

1 Project address

        Project address ：GitHub - open-mmlab/mmclassification: OpenMMLab Image Classification Toolbox and Benchmark

        Installation guide ：

        Prerequisites — MMClassification 0.23.1 documentation

        Getting Started — MMClassification 0.23.1 documentation

2 Installation steps

        mmclassification The installation steps are as follows ：

conda create -n openmmlab python=3.7 -y
conda activate openmmlab
pip install torch==1.8.1+cu101 torchvision==0.9.1+cu101 torchaudio==0.8.1 -f https://download.pytorch.org/whl/torch_stable.html
pip3 install openmim
mim install mmcv-full
git clone https://github.com/open-mmlab/mmclassification.git
# If github Can't connect , You can use gittee Replace .
#git clone https://gitee.com/monkeycc/mmclassification.git
cd mmclassification
pip install -v -e .

        among ,pytorch The version is based on its own cuda Version install the corresponding version . If git Can't use , You can also use the browser to download and unzip . If Github cannot access , You can use gittee Replace , namely ：https://gitee.com/monkeycc/mmclassification.git.

        The installation process prompts are shown in the following figure , Finally, you will be prompted that the installation is successful .

3 Effect test

        （1） Download pre training model

         Run the following command to download the pre training model ：

mim download mmcls --config resnet50_8xb32_in1k --dest .

         After running , Two files will be downloaded in the current directory , namely resnet50_8xb32_in1k.py and resnet50_8xb32_in1k_20210831-ea4938fc.pth.

        If you run the file , Tips search.py Throw the following exception ：

        TypeError: __init__() got an unexpected keyword argument 'autocompletion'

        The reason for this mistake is click Version mismatch , You can reinstall an earlier version of by following the command below click.

 pip install click==7.1.2

        After fixing the above errors ,search.py The file will continue to report the following errors ：

         modify python Environment /lib/python3.7/site-packages/mim/commands/search.py file , This file will also prompt you when you hold it wrong 357 Row threw exception , The error prompt is as follows ：

  File "/root/anaconda3/envs/openmmlab/lib/python3.7/site-packages/mim/commands/search.py", line 169, in get_model_info
    dataframe = convert2df(metadata)
  File "/root/anaconda3/envs/openmmlab/lib/python3.7/site-packages/mim/commands/search.py", line 357, in convert2df
    for result in results:
TypeError: 'NoneType' object is not iterable

            The solution is in 357 Previous increase if results is None: continue. As shown below ：    

if results is None:
    continue
for result in results:

        （2） Run the test script

python demo/image_demo.py demo/demo.JPEG resnet50_8xb32_in1k.py resnet50_8xb32_in1k_20210831-ea4938fc.pth --device cpu

         After running , The following will be output on the command line ：

 4 【python Three dimensional deep learning 】python 3D point cloud from basic to deep learning _Coding Leaves blog -CSDN Blog _python Three dimensional point cloud

More 3D 、 Please pay attention to two-dimensional perception algorithm and financial quantitative analysis algorithm “ Lele perception school ” WeChat official account , And will continue to update .