Kwai-Keye

/

Keye-VL-8B-Preview

like 80

Ask @mradermacher
Request here: https://huggingface.co/mradermacher/model_requests

(They've done llava models before I think, but this is a custom model "Keye" from config.json: mradermacher/llama-joycaption-beta-one-hf-llava-GGUF)

It's not supported by llama.cpp, unfortunately.

Looks like the team is integrating support with vLLM per this PR:
https://github.com/vllm-project/vllm/pull/20126

Might would have success with a bnb 4bit quant or awq after that!