由于直接下载 torch2.1.0，torchvision0.16.0 已不可得，只能手动下载
可以去该网址下载对应版本：https://download.pytorch.org/whl/torch_stable.html
torch下载地址：https://download.pytorch.org/whl/cu121/torch-2.1.0%2Bcu121-cp310-cp310-linux_x86_64.whl
torchvision下载地址：https://download.pytorch.org/whl/cu121/torchvision-0.16.0%2Bcu121-cp310-cp310-linux_x86_64.whl
torchaudio下载地址：https://download.pytorch.org/whl/cu121/torchaudio-2.1.0%2Bcu121-cp310-cp310-linux_x86_64.whl
# 克隆代码
git clone https://github.com/hmkx/HiNeRV.git
cd HiNeRV
# 创建环境
conda create -n hinerv python=3.10
conda activate hinerv
# 安装torch
# torch, torchvision, torchaudio直接下载不可用，需要手动下载文件并安装
pip install torch-2.1.0+cu121-cp310-cp310-linux_x86_64.whl
pip install torchvision-0.16.0+cu121-cp310-cp310-linux_x86_64.whl
pip install torchaudio-2.1.0+cu121-cp310-cp310-linux_x86_64.whl
# 安装依赖
pip install accelerate==0.23.0 -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install deepspeed==0.11.1 -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install pytorch-msssim==1.0.0 -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install timm==0.9.7 -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install torchac==0.9.3 -i https://pypi.tuna.tsinghua.edu.cn/simple
安装 torch, torchvision, torchaudio 的时候可能会报如下错误：
WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'ReadTimeoutError("HTTPSConnectionPool(host='pypi.org', port=443): Read timed out. (read timeout=15)")': /simple/filelock/
原因是：即使下载到本地，然后安装 .whl 文件，pip 仍然会去 pypi.org 下载依赖项（如 filelock、typing_extensions 等），而当前网络无法稳定访问 PyPI，导致超时失败。
解决方案如下：
pip install torch-2.1.0+cu121-cp310-cp310-linux_x86_64.whl -i https://pypi.tuna.tsinghua.edu.cn/simple --trusted-host pypi.tuna.tsinghua.edu.cn
pip install torchvision-0.16.0+cu121-cp310-cp310-linux_x86_64.whl -i https://pypi.tuna.tsinghua.edu.cn/simple --trusted-host pypi.tuna.tsinghua.edu.cn
pip install torchaudio-2.1.0+cu121-cp310-cp310-linux_x86_64.whl -i https://pypi.tuna.tsinghua.edu.cn/simple --trusted-host pypi.tuna.tsinghua.edu.cn
安装好 torch 之后，继续安装其他依赖
安装好其他依赖之后，import torch时可能会有报错：
A module that was compiled using NumPy 1.x cannot be run in
NumPy 2.2.6 as it may crash. To support both 1.x and 2.x
versions of NumPy, modules must be compiled with NumPy 2.0.
Some module may need to rebuild instead e.g. with 'pybind11>=2.12'.If you are a user of the module, the easiest solution will be to
downgrade to 'numpy<2' or try to upgrade the affected module.
We expect that some modules will need time to support NumPy 2.
使用如下方法解决：
pip uninstall numpy
pip install numpy==1.26.4 -i https://pypi.tuna.tsinghua.edu.cn/simple
Training
# 对显存需求极大，需要多卡训练
bash train_hinerv.sh
Evaluation
bash eval_hinerv.sh