mradermacher

/

model_requests

like 93

I'm so excited about dots.llm.1 and I already informed @mradermacher about it in https://huggingface.co/mradermacher/BabyHercules-4x150M-GGUF/discussions/6#684e9961e30cadf2f6406b44 but he seems to be so busy he doesn't even respond there anymore. Let's hope @mradermacher can update to latest llama.cpp soon. I already merged the latest changes to our llama.cpp fork but unfortunately only @mradermacher is able to update all the workers to the latest llama.cpp version.

good luck with the update :)

I do read things, eventually :) I also tend to be excited only after I tried something :) But yeah, it's certainly an exciting model. Maybe I can run a Q4 on my cpu for blazing fast speed...

Q4 is already available, but other quants are not :)

Arruhm... uh... ehm... only an imatrix q4 will do, yes.

As a quick status update: We sucessfully convearted dots.llm1.inst to GGUF and computed its imatrix. We are now computing its static quants following by its imatrix quants. dots.llm1.base is on good tracks as well.

Please monitor https://hf.tst.eu/status.html to always get the latest updates or check the status or regularly check the model
summary page at https://hf.tst.eu/model#dots.llm1.inst-GGUF and https://hf.tst.eu/model#dots.llm1.base-GGUF for quants to appear.

great news!