Update llama4-Scout for AMD

[haic0]

No description provided.

[Richardczl98]

1. Please remove the useless code at the beginning of the document;
2. Please add the Interactive Command Generator;
3. Please add the benchmark;
4. Most importantly, please standardize the format of your document with reference to https://github.com/sgl-project/sgl-cookbook/blob/main/docs/autoregressive/DeepSeek/DeepSeek-V3.md

[haic0]

The following changes have been made, pls check, thanks! 1. 1.the useless code at the beginning of the document removed;
2.the Interactive Command Generator added;
3. the benchmark added;

[jiapingW]

Your docs is about Llama4-Scout but also introduced Llama 4 Maverick. Maybe you can change the docs to Llama4. And use Llama4-Scout as the example. I hope you can fix below content:

• Del port and host in command generator. And add Llama4 Maverick generator too.
• 4.2.1 miss label and the its content is about install sglang which need to placed in Section2.
• Add a accuracy benchmark such as MMLU.
• check the content format and delete unrelated content.

Thanks.

[haic0]

Add Llama-4-Maverick and reformatting the content, pls check, tks.

[jiapingW]

Add Llama-4-Maverick and reformatting the content, pls check, tks.

I see that the command generator only support Llama4-Scout. Can you reformat this PR to Llama4 pr rather than llama4-Scout PR? And check the pr content with other amd's pr to align the format. Thanks.

[haic0]

@jiapingW Thanks, the Llama 4-Maverick added, pls check, thanks!