Add AMD GPU support for Nemotron-Nano-12B-v2-VL

[jiacao-amd]

No description provided.

[gemini-code-assist]

Summary of Changes

Hello @jiacao-amd, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request enhances the existing documentation for the Nemotron-Nano-12B-v2-VL model by integrating a new guide for users with AMD GPUs. The added content provides a step-by-step process, from preparing the Docker environment with ROCm to configuring and running the vLLM server and executing benchmarks, thereby broadening the model's accessibility to a wider range of hardware platforms.

Highlights

• AMD GPU Support Documentation: Added a comprehensive section detailing how to set up and run the Nemotron-Nano-12B-v2-VL model on AMD MI300X and MI355 GPUs.
• Docker Environment Setup: Provided instructions for pulling and launching the vllm/vllm-openai-rocm:v0.14.1 Docker image, including necessary docker run parameters for ROCm compatibility.
• vLLM Server Configuration: Included commands to log in to Hugging Face and start the vLLM online serving with specific environment variables (SAFETENSORS_FAST_GPU=1, VLLM_USE_V1=1, VLLM_USE_TRITON_FLASH_ATTN=0) and model parameters.
• Benchmark Execution: Detailed steps for running a benchmark script inside the Docker container using vllm bench serve with specified model, dataset, input/output lengths, and request rates.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request adds valuable documentation for running Nemotron-Nano-12B-v2-VL on AMD GPUs. The instructions are well-structured. However, I've identified a critical security issue in the provided docker run command that needs to be addressed immediately. I've also pointed out a likely incorrect Docker image version that would prevent users from following the guide, and a point of confusion regarding the number of GPUs assumed in the example commands. My feedback aims to improve the security and clarity of the instructions.

[gemini-code-assist]

Mounting the host's root filesystem (/) into the container with read-write access is extremely dangerous and presents a major security risk. It gives the container unrestricted access to the entire host system. Please change this to mount only a specific, necessary directory. It would also be beneficial to add a security warning for users regarding the use of --privileged and other powerful flags in the docker run command.

Suggested change

	-v /:/work \
	-v <path_to_your_project>:/work \

[gemini-code-assist]

The instruction 'Pull the latest vllm docker:' is contradicted by the command which pins version v0.14.1. This specific version tag also appears to be invalid for the vllm/vllm-openai-rocm image, which will likely cause the setup to fail. Please verify the correct, working version tag and update both the instruction text and the command. If a specific version is required, the text should state that instead of 'latest'.

[gemini-code-assist]

The sample command uses --tensor-parallel-size 8, which assumes an 8-GPU setup. This might be confusing for users, as the preceding text doesn't mention this requirement. Please add a note to clarify that this value should be adjusted based on the user's number of available GPUs (e.g., --tensor-parallel-size 1 for a single GPU).