Update PLaMo 2 Translate serve command

[haic0]

Summary

• Adjust PLaMo 2 Translate to the provided trust-remote-code and TP=4 launch command.
• Aligns the recipe launch guidance with the provided vLLM serve command.

Test plan

• Ran node scripts/build-recipes-api.mjs on the complete validated recipe update set.

Made with Cursor

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
vllm-recipes	Ready	Preview, Comment	Jun 29, 2026 1:52pm

[gemini-code-assist]

Code Review

This pull request updates the configuration and guide for the plamo-2-translate model to default to a tensor parallelism of 4. However, the reviewer notes that hardcoding tp: 4 in strategy_overrides breaks dynamic tensor parallelism scaling for single-GPU setups and recommends removing this override to allow the TP size to adapt dynamically to the selected hardware.

[gemini-code-assist]

Hardcoding tp: 4 in strategy_overrides forces the interactive CommandBuilder to always generate --tensor-parallel-size 4, even when a user selects a single-GPU hardware profile (such as an RTX 4090 or L40S) which is explicitly supported according to the guide. This breaks the dynamic tensor parallelism scaling of the single_node_tp strategy (which normally scales TP to match the selected hardware's GPU count). Removing this override allows the TP size to dynamically adapt to the selected hardware profile.

strategy_overrides: {}