郭宇翔

http和https，具体看对应服务提供商 - DNSPod 的 [移动解析服务D+](https://www.dnspod.cn/httpdns) - AliYun HTTPDNS [HTTPDNS API](https://help.aliyun.com/document_detail/dpa/sdk/RESTful/httpdns.html?spm=5176.docdpa/sdk/OneSDK/quick-start-ios.6.104.wmIJqo) - Google DNS-over-HTTP [介绍](https://developers.google.com/speed/public-dns/docs/dns-over-https)

@deep1401 Is there so way I can workaround? By edit the source code and hard code or provide dtype to start VLLM ?

@deep1401 use FastChat return error "Failed to start model:AssertionError" log: ``` 🏃 Loading Inference Server for deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B with {'inferenceEngine': 'fastchat_server', 'inferenceEngineFriendlyName': ''} Starting Standard FastChat Worker 2025-04-24 14:32:31 | INFO...

@deep1401 No more dtype error, but still can't start server. no detail error, just an `AssertionError` ``` 🏃 Loading Inference Server for Qwen/Qwen3-0.6B-Base with {'inferenceEngine': 'fastchat_server', 'gpu_ids': '', 'model_dtype': 'float16',...

@deep1401 there was not file model_worker_xxx.log when I got error 

> [@deep1401](https://github.com/deep1401) there was not file model_worker_xxx.log when I got error > >  I found some log on `.transformerlab/workspace/logs/model_worker.log` ``` 2025-05-09 11:55:35 | INFO | model_worker | args: Namespace(host='localhost',...

@deep1401 No, it still wan't work. I rm ~/..transformerlab and reinstall ues `curl https://raw.githubusercontent.com/transformerlab/transformerlab-api/main/install.sh | bash` Then when I run a model wtih FastChatServer, it still get AssertionError  ```...

The GPU is too old for this. Thank you very much. @deep1401 @dadmobile

因为我在代码是通过 export default 的方式导出，但是又是通过 ts 编译成了 commonjs，所以用的时候就需要加 default。后续我可以优化一下。 https://remarkablemark.org/blog/2020/05/05/typescript-export-commonjs-es6-modules/