https://huggingface.co/yuhuili/EAGLE3-LLaMA3.3-Instruct-70B
https://huggingface.co/yuhuili/EAGLE3-LLaMA3.3-Instruct-70B
This is another EAGLE3 speculative decoding model that I see quite a bit of sense to GGUF. There are many community fine-tunes of L3.x 70B floating around (and often still in use AFAIK), and speeding these up would be nice, I think.
Many thanks!
(RedHatAI also have eagle3 for L 3.3 70B Instruct, which was not mentioned in the freshly merged llama.cpp PR however; I point at eagle3 model, which PR specifically mentions support for)
(for sake of completeness, I refer to same PR as prior, this one ~ https://github.com/ggml-org/llama.cpp/pull/18039 )
same as last one =)
It's queued!
You can check for progress at http://hf.tst.eu/status.html or regularly check the model
summary page at https://hf.tst.eu/model#EAGLE3-LLaMA3.3-Instruct-70B-GGUF for quants to appear.