https://huggingface.co/RedHatAI/gemma-4-31B-it-speculator.eagle3
https://huggingface.co/RedHatAI/gemma-4-31B-it-speculator.eagle3
Support for EAGLE3 speculative decoding models was recently added to llama.cpp; there is quite many of these eagle3 models, and I think that it is reasonable to start GGUF'ing with this one, since it is directly mentioned in the merged PR, and is also of a fresh good non-MoE (gets more speedup, to my understanding) model.
Thank you for the wonderful quantization work you do!
And in case you are very curious to know more of these eagle3 things...
Here's the PR ~ https://github.com/ggml-org/llama.cpp/pull/18039
And here's a reddit discussion (I am happy that the old design of reddit is still an option) ~ https://old.reddit.com/r/LocalLLaMA/comments/1u3on4u/eagle3_has_landed_in_llamacpp/
It's queued! Im not a public reddit user, but I updated llama cpp, and if it works you can share the link for our quant =)
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#gemma-4-31B-it-speculator.eagle3-GGUF for quants to appear.