Skip to content

[New Model]: DeepSeek-VL2 #11236

Closed
Closed
@mgoin

Description

@mgoin

The model to consider.

https://huggingface.co/deepseek-ai/deepseek-vl2
https://huggingface.co/deepseek-ai/deepseek-vl2-small
https://huggingface.co/deepseek-ai/deepseek-vl2-tiny

The closest model vllm already supports.

DeepSeekV2 is the base language model, so that should already be supported. From what I can tell the new vision support is simply siglip with an mlp projector. https://huggingface.co/deepseek-ai/deepseek-vl2/blob/e6adb2bce35b94ecc84fbb46d130ce60a7bb4d43/config.json#L129-L144

What's your difficulty of supporting the model you want?

I think should be similar to supporting any other Llava-style model, we should have all the pieces implemented and just need a new model definition. However support has not landed in transformers yet.

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

No one assigned

    Labels

    new-modelRequests to new models

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions