mrfakename/Beam-1.0-REAP-255B-A23B
#3015
by tomasmcm - opened
https://huggingface.co/mrfakename/Beam-1.0-REAP-255B-A23B
"50% expert-pruned" should allow running this under 128GB RAM 🙏
It's queued! but automatically classified as big, let's hope someone manages to get it manually done... and let's hope llamacpp agrees with 50% experts being pruned and wont search for them lol
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#Beam-1.0-REAP-255B-A23B-GGUF for quants to appear.
let's hope llamacpp agrees with 50% experts being pruned and wont search for them lol
Is that a thing? 😅 maybe I've been lucky with the reaps I've tried so far.
Yeah the model is quite big, hence why a reap would be much more practical for people to try