fix dummy-load deepseek nextn

[lambert0312]

Motivation

Ref #4535

Modifications

Purely load logic for deepseek_nextn.py.

Checklist

• Format your code according to the Code Formatting with Pre-Commit.
• Add unit tests as outlined in the Running Unit Tests.
• Update documentation / docstrings / example tutorials as needed, according to Writing Documentation.
• Provide throughput / latency benchmark results and accuracy evaluation results as needed, according to Benchmark and Profiling and Accuracy Results.
• For reviewers: If you haven't made any contributions to this PR and are only assisting with merging the main branch, please remove yourself as a co-author when merging the PR.
• Please feel free to join our Slack channel at https://slack.sglang.ai to discuss your PR.

[ispobock]

Could you add accuracy test result?

[lambert0312]

Could you add accuracy test result?

@ispobock The accuracy test results are as follows:

• gsm8k

Accuracy: 0.950
Invalid: 0.000
Latency: 14.089 s
Output throughput: 1527.533 token/s

• mmlu

Total latency: 146.207
Average accuracy: 0.875

[ispobock]

I think we can merge this PR first. I have some refactor for the nextn part recently, may have some conflict with this PR. I can rebase it after it merged.
cc: @merrymercy

[lambert0312]

I think we can merge this PR first. I have some refactor for the nextn part recently, may have some conflict with this PR. I can rebase it after it merged. cc: @merrymercy

See if it can be merged? @merrymercy @ispobock @zhyncs

[lambert0312]

@ispobock I see your PR #5793 has been merged. My PR seems unnecessary. I will close it.