mradermacher

/

GLM-4.5-Air-i1-GGUF

like 3

Me, too, unfortunately, llama.cpp currently fails to generate them :(

As mentioned in https://huggingface.co/mradermacher/model_requests/discussions/1239#6891b806a57cc4b88f6d0437 there unfortunately won't be any low bits-per-wight quants for any GLM 4.5 based models from us until MTP support is implemented by llama.cpp after which we will have to redo all the GLM 4.5 based models anyways.

@mradermacher When we are at GLM 4.5 please mark GLM-4.5 and GLM-4.5-Base to use the imatrix RPC setup which I will fix as soon I feel better again.

@nicoboss marked

Thank you Señors.
I will be looking forward for the update in the future.

Best, Pilo.