base_model

base_model
Qwen/Qwen2.5-VL-7B-Instruct

NOTE: OLDER / OTHER VERSIONS OF THIS MODEL ARE ERRONEOUSLY PRODUCED Qwen2VLxQwen2.5 HYBRIDS!!!

Although these models have seemingly better performance than Qwen2VL, they are not the real deal! True Qwen2.5VL GGML support is still being added.

To test it out:

https://github.com/Independent-AI-Labs/llama.cpp/commits/debug/build/

You will have to build llama.cpp with CLIP hardware acceleration enabled to get the new architecture running.

Monitor our progress in merging this upstream here: https://github.com/ggml-org/llama.cpp/pull/12119

Official release and benchmarks coming soon!