UDIS2复现
1.将文件environment.yml放在C:\Users\sun>文件夹下
2.conda env create -f environment.yml
environment.yml中的软件包是为Linux构建的，无法在Windows上安装
以上方法在笔者的现有的条件下行不通，换一种方式
Requirement：
numpy 1.19.5
pytorch 1.7.1
scikit-image 0.15.0
tensorboard 2.9.0
1.conda create -n py3.8 python=3.8.0
2.conda install numpy=1.19.5 scikit-image=0.15.0 tensorboard=2.9.0 pytorch=1.7.1 cudatoolkit=11.0 -c pytorch
原环境配置中包含torchvision=0.8.2和torchaudio=0.7.2，这些通常与PyTorch一起使用，尤其在图像处理任务中
3.conda install torchvision=0.8.2 torchaudio=0.7.2
环境暂时完成，可能存在一些问题，先走入下一步，遇到问题再解决
运行时发现缺少环境：
1.pip install opencv-python -i https://pypi.tuna.tsinghua.edu.cn/simple
2.conda install imageio
3.conda install scikit-image
运行失败，删除环境
新建环境，再换一种方式
可以配置 TUNA 清华镜像源，速度更快，且避开代理问题：
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/r/
conda config --set show_channel_urls yes
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
conda config --add channels defaults
1.基础依赖和开发工具
conda install _anaconda_depends=2020.07=py38_0 _ipyw_jlab_nb_ext_conf=0.1.0=py38_0 alabaster=0.7.12 anaconda=custom=py38_1 anaconda-client=1.7.2 anaconda-navigator=1.10.0 anaconda-project=0.8.4 argh=0.26.2 argon2-cffi=20.1.0 asn1crypto=1.4.0 astroid=2.4.2 astropy=4.0.2 async_generator=1.10 atomicwrites=1.4.0 attrs=20.3.0 autopep8=1.5.4 babel=2.8.1 backcall=0.2.0 backports=1.0 backports.functools_lru_cache=1.6.1 backports.shutil_get_terminal_size=1.0.0 backports.tempfile=1.0 backports.weakref=1.0.post1 beautifulsoup4=4.9.3 bitarray=1.6.1 bkcharts=0.2 blas=1.0=mkl bleach=3.2.1 blosc=1.20.1 bokeh=2.2.3 boto=2.49.0 bottleneck=1.3.2 brotlipy=0.7.0 bzip2=1.0.8
2.数据科学和可视化
conda install cffi=1.14.3 chardet=3.0.4 click=7.1.2 cloudpickle=1.6.0 clyent=1.2.2 colorama=0.4.4 conda-package-handling=1.7.2 conda-verify=3.4.2 contextlib2=0.6.0.post1 cryptography=3.1.1 cudatoolkit=11.0.221 curl=7.71.1 cycler=0.10.0 cython=0.29.21 cytoolz=0.11.0 dask=2.30.0 dask-core=2.30.0 decorator=4.4.2 defusedxml=0.6.0 distributed=2.30.1 docutils=0.16 entrypoints=0.3 et_xmlfile=1.0.1 expat=2.2.10 fastcache=1.1.0 filelock=3.0.12 flake8=3.8.4 flask=1.1.2 fontconfig=2.13.0 freetype=2.10.4
没有fontconfig=2.13.0，先不安装这个包
有一些包被安装，有一些包被被优先级更高的通道取代
3.科学计算、NLP、Notebook 相关
conda install future=0.18.2 get_terminal_size=1.0.0 gevent=20.9.0 glob2=0.7 greenlet=0.4.17 h5py=2.10.0 hdf5=1.10.4 heapdict=1.0.1 html5lib=1.1 icu=58.2 idna=2.10 imageio=2.9.0 imagesize=1.2.0 importlib_metadata=2.0.0 iniconfig=1.1.1 intel-openmp=2020.2 intervaltree=3.1.0 ipykernel=5.3.4 ipython=7.19.0 ipython_genutils=0.2.0 ipywidgets=7.5.1 isort=5.6.4 itsdangerous=1.1.0 jedi=0.17.1 jinja2=2.11.2 joblib=0.17.0 json5=0.9.5 jsonschema=3.2.0 jupyter=1.0.0 jupyter_client=6.1.7 jupyter_console=6.2.0 jupyter_core=4.6.3 jupyterlab=2.2.6 jupyterlab_pygments=0.1.2 jupyterlab_server=1.2.0 keyring=21.4.0 kiwisolver=1.3.0
有一些包被移除，有一些包被更新，有一些包被被优先级更高的通道取代，有一些包被降级
4.数值计算、Numpy、Pandas
conda install krb5=1.18.2 lazy-object-proxy=1.4.3 libarchive=3.4.2 libcurl=7.71.1 libedit=3.1.20191231 libffi=3.3 libpng=1.6.37 libtiff=4.1.0 libuuid=1.0.3 libuv=1.40.0 libxml2=2.9.10 libxslt=1.1.34 llvmlite=0.34.0 locket=0.2.0 lxml=4.6.1 lz4-c=1.9.2 lzo=2.10 markupsafe=1.1.1 matplotlib=3.3.2 matplotlib-base=3.3.2 mccabe=0.6.1 mistune=0.8.4 mkl=2020.2 mkl-service=2.3.0 mkl_fft=1.2.0 mkl_random=1.1.1 mock=4.0.2 more-itertools=8.6.0 mpmath=1.1.0 msgpack-python=1.0.0 multipledispatch=0.6.0 nbclient=0.5.1 nbconvert=6.0.7 nbformat=5.0.8 networkx=2.5 ninja=1.10.2 nltk=3.5 nose=1.3.7 notebook=6.1.4 numba=0.51.2 numexpr=2.7.1 numpy-base=1.19.2 numpydoc=1.1.0
没有libuuid=1.0.3，libffi=3.3，libedit=3.1.20191231，先不安装这些包
有一些包被安装，有一些包被被优先级更高的通道取代，有一些包被降级
运行时提示没有torch，安装一下conda install pytorch=1.7.1 -c pytorch
运行时提示没有tensorboard，安装一下conda install tensorboard=2.9.0
运行时提示没有cv2，安装一下pip install opencv-python -i https://pypi.tuna.tsinghua.edu.cn/simple
运行时提示没有torchvision，安装一下conda install torchvision=0.8.2
貌似要成功了，继续调试一下
1.下载数据集
2.电脑内存不够，使用虚拟内存（硬盘的一部分容量），硬盘的读写速度远低于内存，虚拟内存的大小也影响电脑运行的流畅性和稳定性
3.GPU显存不够
4.使用CPU跑
D:\Anaconda_python3.12\envs\py3.8\python.exe D:\desktop\learn_dl\UDIS2-main\Warp\Codes\train.py
<==================== setting arguments ===================>
Namespace(batch_size=1, gpu='0', max_epoch=100, train_path='D:\\desktop\\learn_dl\\UDIS2-main\\Warp\\Datasets\\training')
odict_keys(['input1', 'input2'])
training from stratch!
##################start training#######################
start epoch 0
0 lr=0.000100
Traceback (most recent call last):
File "D:\desktop\learn_dl\UDIS2-main\Warp\Codes\train.py", line 181, in <module>
train(args)
File "D:\desktop\learn_dl\UDIS2-main\Warp\Codes\train.py", line 100, in train
batch_out = build_model(net, inpu1_tesnor, inpu2_tesnor)
File "D:\desktop\learn_dl\UDIS2-main\Warp\Codes\network.py", line 137, in build_model
H_motion, mesh_motion = net(aug_input1_tensor, aug_input2_tensor)
File "D:\Anaconda_python3.12\envs\py3.8\lib\site-packages\torch\nn\modules\module.py", line 727, in _call_impl
result = self.forward(*input, **kwargs)
File "D:\desktop\learn_dl\UDIS2-main\Warp\Codes\network.py", line 530, in forward
correlation_64 = self.CCL(feature_1_64, warp_feature_2_64)
File "D:\desktop\learn_dl\UDIS2-main\Warp\Codes\network.py", line 607, in CCL
flow_w = match_vol*(c_one%w - w_one)
RuntimeError: [enforce fail at ..\c10\core\CPUAllocator.cpp:73] data. DefaultCPUAllocator: not enough memory: you tried to allocate 67108864 bytes. Buy new RAM!
进程已结束，退出代码为 1
无法打开tensorboard
(base) D:\desktop\learn_dl\UDIS2-main\Warp>tensorboard --logdir=summary
Traceback (most recent call last):
File "D:\Anaconda_python3.12\envs\py3.8\Scripts\tensorboard-script.py", line 6, in <module>
from tensorboard.main import run_main
File "D:\Anaconda_python3.12\envs\py3.8\lib\site-packages\tensorboard\main.py", line 27, in <module>
from tensorboard import default
File "D:\Anaconda_python3.12\envs\py3.8\lib\site-packages\tensorboard\default.py", line 30, in <module>
import pkg_resources
File "D:\Anaconda_python3.12\envs\py3.8\lib\site-packages\pkg_resources\__init__.py", line 95, in <module>
from jaraco.text import drop_comment, join_continuation, yield_lines
File "D:\Anaconda_python3.12\envs\py3.8\lib\site-packages\setuptools\_vendor\jaraco\text\__init__.py", line 12, in <module>
from jaraco.context import ExceptionTrap
File "D:\Anaconda_python3.12\envs\py3.8\lib\site-packages\setuptools\_vendor\jaraco\context.py", line 17, in <module>
from backports import tarfile
ImportError: cannot import name 'tarfile' from 'backports' (D:\Anaconda_python3.12\envs\py3.8\lib\site-packages\backports\__init__.py)
pip install setuptools backports.tarfile
显卡比较：Tesla T4 与 RTX3090Ti 性能对比；深度学习方向效率对比；
原始资料地址：
UDIS2
如有侵权联系删除 仅供学习交流使用