Support Qwen VL

[thesby]

Could you please support Qwen VL model

[Boomboomdunce]

Hope to increase the support of Qwen VL

[thiner]

This is really a must have model for Chinese users, as it's the SOTA of vision model which can recognize Chinese characters. But I think that won't be easy, because there is not GGUF format model yet on HuggingFace.

[thesby]

There is a int4 model. https://huggingface.co/Qwen/Qwen-VL-Chat-Int4

[thiner]

@thesby ollama runs GGUF models only currently. Pytorch or safetensors model need to be converted to gguf firstly. https://github.com/ollama/ollama/blob/main/docs/import.md. But convert Qwen-VL to gguf format is not supported by llama.cpp yet.

[monkeycc]

Hope to increase the support of Qwen VL +1

[ghost]

hope + 1

[TKasperczyk]

+1

[leo985]

+1

[wangqinghuan]

+1

[J0eky]

+1

[liyanshuai2018]

This multimodal model is really great, supported, and it is recommended to adapt

[BayMikyou]

+1

[meloseven]

+1

[bobbyng626]

+1

[meloseven]

这是来自QQ邮箱的假期自动回复邮件。您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。

[aboutmydreams]

llama.cpp 尚不支持将 Qwen-VL 转换为 gguf 格式，得去他们仓库下多 dd...

[meloseven]

这是来自QQ邮箱的假期自动回复邮件。您好，我最近正在休假中，无法亲自回复您的邮件。我将在假期结束后，尽快给您回复。

[yjgfn]

+1

[janelovesprogramming]

+1

[chrisoutwright]

Qwen VL works much better than LLava 1.6, would be good to be able to use it with ollama. The OCR is also much better and supports much more languages.