事项要素模型

环境问题

普通Linux环境没什么问题

arm＋麒麟系统

• python3.7
• tensorflow==1.14.0(tensorflow-1.14.0-cp37-none-linux_aarch64.whl)
• keras==2.3.0
• h5py==2.10.0(conda install h5py==2.10.0)
• protobuf==3.20.0

docker操作：docker cp xxx 6a8:/root
docker exec -it 6a8 bash

报错

File "/home/qeh/anaconda3/envs/model-system/lib/python3.7/site-packages/keras/engine/saving.py", line 492, in load_wrapper
    return load_function(*args, **kwargs)
  File "/home/qeh/anaconda3/envs/model-system/lib/python3.7/site-packages/keras/engine/network.py", line 1230, in load_weights
    f, self.layers, reshape=reshape)
  File "/home/qeh/anaconda3/envs/model-system/lib/python3.7/site-packages/keras/engine/saving.py", line 1183, in load_weights_from_hdf5_group
    original_keras_version = f.attrs['keras_version'].decode('utf8')
AttributeError: 'str' object has no attribute 'decode'

• h5py版本问题
• pip install h5py==2.10.0

安装h5py时遇到的问题

PackagesNotFoundError: The following packages are not available from current channels:

• h5py==2.10.0
  Current channels:
• https://repo.anaconda.com/pkgs/main/linux-aarch64
• https://repo.anaconda.com/pkgs/main/noarch
• https://repo.anaconda.com/pkgs/r/linux-aarch64
• https://repo.anaconda.com/pkgs/r/noarch
  To search for alternate channels that may provide the conda package you’re
  looking for, navigate to
  https://anaconda.org
  and use the search bar at the top of the page.

conda config --append channels conda-forge
conda search h5py==2.10.0
conda install h5py==2.10.0

# 不行的话尝试这个
conda install -c conda-forge h5py==2.10.0

安装h5py会遇到各种不可控问题，最好打包conda环境来移植。

conda install conda-pack
# -o 参数指定打包路径和名称
conda pack -n 虚拟环境名 -o 打包环境名.tar.gz
# 创建目录 `my_env`，并将环境解压至该目录
mkdir -p my_env
tar -xzf my_env.tar.gz -C my_env