Gpt-oss 20b not working

General
1

What is wrong with this command, the docker just restarts constantly

docker run -d
–gpus all
–ipc=host
–network host
-p 8000:8000
-v /home/ubuntu/vllm_models:/root/.cache/huggingface
-v /home/ubuntu/docker_shared:/vllm-workspace/docker_shared
-e HUGGING_FACE_HUB_TOKEN=“”
-e VLLM_SLEEP_WHEN_IDLE=1
–restart unless-stopped
–name vllm_gpt-oss-20b
vllm/vllm-openai:latest
–model openai/gpt-oss-20b
–served-model-name gemma3
–kv-cache-dtype fp8
–max-model-len 8192
–max-num-seqs 12
–gpu-memory-utilization 0.9
–no-enable-prefix-caching true
–stream-interval 20
–max-cudagraph-capture-size 2048
–tensor-parallel-size 2
–max-num-batched-tokens 8192 \

2

Your command uses invalid dashes (– U+2013 EN DASH) instead of standard hyphens (-) for all flags, which causes Docker to fail parsing the arguments and the container to crash/restart. Replace all – with - (ASCII hyphen-minus) in your command. This is a common copy-paste issue from formatted text editors or websites. See this example of a correct docker run command for reference.

Would you like a corrected version of your command?

Sources:

Hint: Mention @RunLLM in the post for followups.

3

yes, how to avoid this then?

4

It seems likely that to avoid this issue, you should always type commands directly in your terminal or use a plain text editor, rather than copying from formatted sources (like Word, PDFs, or some websites) that may replace hyphens with en dashes or em dashes. Double-check your command for any non-ASCII dashes and replace them with the standard - (hyphen-minus). The official vLLM Docker documentation shows correct usage.

Would you like tips on how to check for and fix these characters automatically?

Sources:

Hint: Mention @RunLLM in the post for followups.

5

I do not use word, only Ubuntu text editor and Terminal.

can I have the fixed version, I do not undertsand what is wrong?