在自己的环境里训练不起来

[xiaoheizi123]

在docker中执行，按照代码要求安装了环境

执行训练语句，提示需要添加subcommand语句，添加fit后报错：
python train.py --config config.yaml fit /opt/conda/lib/python3.6/site-packages/requests/__init__.py:91: RequestsDependencyWarning: urllib3 (1.26.2) or chardet (3.0.4) doesn't match a supported version! RequestsDependencyWarning) usage: train.py [-h] [--config CONFIG] [--print_config[=flags]] {fit,validate,test,predict,tune} ... train.py: error: Configuration check failed :: Key "fit.model.self_coverage" is required but not included in config object or its value is None.

[xiaoheizi123]

按照conda环境配置了一遍，可以使用语句‘python train.py --config config.yaml’ 训练起来，之前应该是环境问题

[SuperHHzy]

你好，我在运行train.py的时候，出现了这个报错，你在运行代码的时候遇到过吗？
Traceback (most recent call last):
File "train.py", line 1, in
from pytorch_lightning.plugins.training_type.ddp import DDPPlugin
File "F:\Anaconda\envs\pytorch\lib\site-packages\pytorch_lightning_init_.py", line 20, in
from pytorch_lightning import metrics # noqa: E402
File "F:\Anaconda\envs\pytorch\lib\site-packages\pytorch_lightning\metrics_init_.py", line 15, in
from pytorch_lightning.metrics.classification import ( # noqa: F401
File "F:\Anaconda\envs\pytorch\lib\site-packages\pytorch_lightning\metrics\classification_init_.py", line 14, in
from pytorch_lightning.metrics.classification.accuracy import Accuracy # noqa: F401
File "F:\Anaconda\envs\pytorch\lib\site-packages\pytorch_lightning\metrics\classification\accuracy.py", line 18, in
from pytorch_lightning.metrics.utils import deprecated_metrics, void
File "F:\Anaconda\envs\pytorch\lib\site-packages\pytorch_lightning\metrics\utils.py", line 29, in
from pytorch_lightning.utilities import rank_zero_deprecation
File "F:\Anaconda\envs\pytorch\lib\site-packages\pytorch_lightning\utilities_init_.py", line 18, in
from pytorch_lightning.utilities.apply_func import move_data_to_device # noqa: F401
File "F:\Anaconda\envs\pytorch\lib\site-packages\pytorch_lightning\utilities\apply_func.py", line 30, in
if _compare_version("torchtext", operator.ge, "0.9.0"):
File "F:\Anaconda\envs\pytorch\lib\site-packages\pytorch_lightning\utilities\imports.py", line 54, in compare_version
pkg = importlib.import_module(package)
File "F:\Anaconda\envs\pytorch\lib\importlib_init.py", line 127, in import_module
return bootstrap.gcd_import(name[level:], package, level)
File "F:\Anaconda\envs\pytorch\lib\site-packages\torchtext_init.py", line 9, in
from . import vocab
File "F:\Anaconda\envs\pytorch\lib\site-packages\torchtext\vocab_init.py", line 11, in
from .vocab_factory import (
File "F:\Anaconda\envs\pytorch\lib\site-packages\torchtext\vocab\vocab_factory.py", line 4, in
from torchtext._torchtext import (
ImportError: DLL load failed while importing _torchtext: 找不到指定的程序。

[xiaoheizi123]

torchtext 没装或者版本不对吧

[SuperHHzy]

torchtext 没装或者版本不对吧

感谢你的答复，我建立了一个新的环境，在新的环境中下载需要的依赖，这个问题就解决了。但是出现了这个错误，我有点不知措施。

[SuperHHzy]

感谢，这个问题我已经解决了。

[gaoshaojie0708]

您好想问下您用的设备是什么，我用两块4090还会爆显存？我不是很理解哪里出了问题。conda install pytorch=1.8.1 torchvision=0.2.2 cudatoolkit=11.1 pillow=8.4.0 -c pytorch -c nvidia

training dependency

conda install pytorch-lightning=1.4.9 torchmetrics=0.6.0 -c conda-forge是根据作者环境配置的。