subprocess.CalledProcessError: Command '['python3', '/home/b/hxb/yolox/YOLOX/tools/train.py']' returned non-zero exit status 1.

有人能帮我解决一下问题吗？
我用pycharm链接服务器后，在pycharm上运行train.py会遇到以下问题：
ssh://b@222.197.000.10:9022/home/b/miniconda3/envs/yolo1.7/bin/python -u /home/b/hxb/yolox/YOLOX/tools/train.py
2021-08-03 09:34:38.959 | INFO | yolox.core.launch:launch_by_subprocess:145 -

Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed.

Traceback (most recent call last):
File "/home/b/hxb/yolox/YOLOX/tools/train.py", line 5, in
from loguru import logger
ModuleNotFoundError: No module named 'loguru'
Traceback (most recent call last):
File "/home/b/hxb/yolox/YOLOX/tools/train.py", line 5, in
from loguru import logger
ModuleNotFoundError: No module named 'loguru'
Traceback (most recent call last):
File "/home/b/hxb/yolox/YOLOX/tools/train.py", line 135, in
args=(exp, args),
File "/home/b/hxb/yolox/YOLOX/yolox/core/launch.py", line 87, in launch
args,
File "/home/b/hxb/yolox/YOLOX/yolox/core/launch.py", line 165, in launch_by_subprocess
raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd)
subprocess.CalledProcessError: Command '['python3', '/home/b/hxb/yolox/YOLOX/tools/train.py']' returned non-zero exit status 1.
我尝试了一下，如果在train.py中修改devices参数来只使用单卡的话，便不会报这个错误。但当我使用两张卡时，这个错误便出现了。
如果将from loguru import logger 放在 import torch后面，则会报错：
ssh://b@222.197.219.10:9022/home/b/miniconda3/envs/yolo1.7/bin/python -u /home/b/hxb/yolox/YOLOX/tools/train.py
2021-08-03 09:46:28.464 | INFO | yolox.core.launch:launch_by_subprocess:145 -

Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed.

Traceback (most recent call last):
File "/home/b/hxb/yolox/YOLOX/tools/train.py", line 7, in
import torch
ModuleNotFoundError: No module named 'torch'
Traceback (most recent call last):
File "/home/b/hxb/yolox/YOLOX/tools/train.py", line 7, in
import torch
ModuleNotFoundError: No module named 'torch'
Traceback (most recent call last):
File "/home/b/hxb/yolox/YOLOX/tools/train.py", line 135, in
args=(exp, args),
File "/home/b/hxb/yolox/YOLOX/yolox/core/launch.py", line 87, in launch
args,
File "/home/b/hxb/yolox/YOLOX/yolox/core/launch.py", line 165, in launch_by_subprocess
raise subprocess.CalledProcessError(returncode=process.returncode, cmd=cmd)
subprocess.CalledProcessError: Command '['python3', '/home/b/hxb/yolox/YOLOX/tools/train.py']' returned non-zero exit status 1.
运行的代码是昨天下刚载的。
如果直接用Xshell在服务器上运行train.py则没有这个问题。

我的环境：
CUDA10.1 虚拟环境：python 3.7 pytorch==1.7.0 torchvision==0.8.0 torchaudio==0.7.0 cudatoolkit=10.1

希望有人能帮我解答一下，谢谢！

Start with tools/train.py and yolox/core/launch.py, especially launch_by_subprocess, to trace how the multi-GPU workers are started. Compare the interpreter and environment used by the subprocess with the PyCharm and Xshell environments, checking imports for loguru and torch. Done means multi-GPU training starts with the intended environment without ModuleNotFoundError.