Skip to content

[Feature]: ROCm H3 support #5697

Description

@functionstackx

🚀 The feature, motivation and pitch

hi @andyluo7 can we get an ROCm MiniMax H3 support in vllm-omni. seems like quite an popular model

currently it is only supported on CUDA https://recipes.vllm.ai/MiniMaxAI/MiniMax-H3

Alternatives

No response

Additional context

No response

Before submitting a new issue...

  • Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

Metadata

Metadata

Assignees

Labels

ROCmPR related to AMD hardwarediffusioncodes related to diffusion models

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions