baymax55

> It's supposed to be `from revChatGPT.revChatGPT import Chatbot` well, I do not know why it does not work, just using local import is fine 

> > > It's supposed to be `from revChatGPT.revChatGPT import Chatbot` > > > > > > well, I do not know why it does not work, just using local...

> It's supposed to be `from revChatGPT.revChatGPT import Chatbot` I finally found move code to the src directory is fine, thank you 

这只是临时的方案,2小时后token依旧会过期.仍需手动续期 https://github.com/acheong08/ChatGPT/wiki/Setup#session-token-authentication ` Use this only if all else fails. Refreshing the session does not work here. You have to refresh manually.`

> Thanks @baymax55 for this PR. Would be great if you also could add a simple test for this new API - there should be plenty of samples in that...

Server 默认是支持集群部署吧,通过nginx反向代理后面的server集群,集群共用一个数据库实例.

> @baymax55 有部分数据是在内存中的，无法正常按照多节点集群部署。需要让内存中的数据在集群中交互。 可以举例说下是哪部分数据存到内存中了，我觉得，将内存数据抽回来会比让server实例同步数据好些

I also encountered this problem. `Error from server (BadRequest): error when creating "crd.yml": CustomResourceDefinition in version "v1" cannot be handled as a CustomResourceDefinition: strict decoding error: unknown field "spec.versions[0].schema.properties", unknown...

构建Dockerfile.gpu出错时,将下列代码进行替换: ` RUN cat > /get-models.py > /get-models.py && \ echo "import torch" >> /get-models.py && \ echo "config = Config()" >> /get-models.py && \ echo "config.device = 'cuda' if...