I'm on waiting list for AesSedai Minimax M2.7 Q4_K_M or similar...
#3
by MartinPatterson - opened
If and when it comes...
Cheers!
As soon as it drops and is supported in llama.cpp (hopefully it's the same arch), I'll have the quants available soon after :)
Very need iq4_xs with same size as minimax m2.5 iq4_xs. Thank you!!
I've got the weights downloaded, working on them tonight!
Please, pretty please don't skip the "Q5_K_M" more so now when mainline has decided to finally use CUDA for the Q5 quants and not offload it to the CPU anymore (since yesterday!). Thank you!