此电脑右键--属性高级系统设置-- 环境变量--系统变量里找到 Path 然后选择编辑进入下面的页面后，选择新建，然后把你装的 python 路径放上然后测试，在一个写好测试文件 (hello.py) 的文件夹里按住shift 右键，选择在此处打开 Powershell 窗口输入 python hello.py...

1. 代码连接# 运行后出现URL，点击获得你的认证码from google.colab import drivedrive.mount('/content/drive')# 列出你云盘里有啥! ls "/content/drive//My Drive/"# 读取云盘里的文件喽，我这里读的是 bbc-text.csvwith open("/content/drive//My Drive/bbc-t

本地加载roberta-base模型文件，在该网站下载模型文件：roberta-base at main (huggingface.co)所需的有 config.json, merges.txt, pytorch_model.bin, vocab.json路径组织结构：└─model│ merges.txt│ vocab.json│└─roberta-baseconfig.json...

进入 hugging face 的这个网站：https://huggingface.co/bert-base-chinese#如下图，点开 list files in model 就可以看到需要下载的文件了

有点匪夷所思，错误解决是在 model 模块，# 报错logits = self.linear(pooled_output)# 没问题了logits = self.linear(pooled_output).view(batch_size, self.num_classes)

https://github.com/pytorch/pytorch/issues/35803pip install intel-openmp

用的多GPU的数据并行方法DataParallel ，这老出错原 batch_size 我设的 8，谷歌到该 github issuehttps://github.com/Eromera/erfnet_pytorch/issues/2然后 batch_size 设为 9 目前能跑通

vscode could not establish connection to linux The VS Code Server failed to startThe remote host may not meet VS Code Server's prerequisites for glibc and libstdc++

vscode could not establish connection to linux The VS Code Server failed to startThe remote host may not meet VS Code Server's prerequisites for glibc and libstdc++