Issue with run gpt-oss-120b in vLLM

[AastaLLL]

Hi,

The configuration should be similar to the standard vLLM usage.
As @cailuyu suggested, please try if --max-model-len helps for your requirement.

Thanks.

[johnny_nv]

Hello,
I was working on VLLM repository.
If you build vllm now, it is working fine for Thor

[cailuyu]

yes ,without backend setting, it will be ok！