踩坑记录yolov5环境版本要求比较严苛
目录
踩坑记录：yolov5环境版本要求比较严苛？
在安装yolov5环境时，numpy安装失败报错metadata-generation-failed
报错如下：
Collecting numpy==1.18.5 (from -r /*****/yolov5-5.0/requirements.txt (line 5))
Using cached https://pypi.tuna.tsinghua.edu.cn/packages/01/1b/d3ddcabd5817be02df0e6ee20d64f77ff6d0d97f83b77f65e98c8a651981/numpy-1.18.5.zip (5.4 MB)
......
replace_method(CCompiler, 'find_executables', CCompiler_find_executables)
NameError: name 'CCompiler' is not defined
[end of output]
note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed
× Encountered error while generating package metadata.
╰─> See above for output.
note: This is an issue with the package mentioned above, not pip.
hint: See above for details.在python 3.7，3.8，3.9，3.10，3.12均无法正常安装numpy，或报错要求更高python版本。
遂调整requirements.txt的版本如下：
注：python要求3.8版本，否则torch可能无法安装1.7.1。
# pip install -r requirements.txt
# base ----------------------------------------
matplotlib==3.5.0
numpy==1.21.4
#numpy==1.18.5
opencv-python==4.5.4.60
Pillow==8.4.0
PyYAML==6.0
scipy==1.7.3
torch==1.7.1
#torch==1.7.0
torchvision>=0.8.1
tqdm==4.41.0
# logging -------------------------------------
tensorboard==2.8.0
# wandb
# plotting ------------------------------------
seaborn==0.11.2
pandas==1.3.4
# export --------------------------------------
# coremltools==4.1
# onnx==1.8.1
# scikit-learn==0.19.2 # for coreml quantization
# extras --------------------------------------
thop # FLOPS computation
pycocotools==2.0 # COCO mAP发现torch安装失败，遂使用conda进行安装：
conda install pytorch==1.11.0 torchvision==0.12.0 torchaudio==0.11.0 cudatoolkit=11.3 -c pytorch此命令，conda将自动拉取所需依赖，并在虚拟环境下安装cuda，和torch相关库。
注：numpy被指定版本1.21.4， scipy被指定版本1.7.3
报错：
ModuleNotFoundError: No module named 'Cython'
No matching distribution found for pycocotools==2.0安装对应库：
pip install Cython pycocotools==2.0