Req: MLX Q4 or similar for 64GB Apple Silicon?
Hi, great work, excited to run this. I'd like to run it on 64GB Apple Silicon with less of a squeeze.
I hope it's ok to ask: would someone please be able to provide a MiniMax-M2.1-PRISM MLX 4bit (Q4_K_M style) build or similar, optimized for Apple Silicon (e.g., LM Studio)?
Thank you.
@JuanPabloski - We'll take a look at how to best provide additional quants while spreading the cost for custom asks. Please consider supporting our work.
Do you accept Bitcoin?
If you wish to sponsor our work Here's my BTC address:
bc1qkhh8k7t4v48g6sr0nxxjpevktkea8vmez97qas
-E.
If you wish to sponsor our work Here's my BTC address:
bc1qkhh8k7t4v48g6sr0nxxjpevktkea8vmez97qas-E.
I've sent something to your wallet, thank you very much for your work. I hope you can upload gguf or mlx quantizations of the glm 4.7 prism model.
Thank you @Asencion — much appreciated! We’re working on something very special. Stay tuned on X and HF for our next major releases.
For GLM-4.7, we’re developing a new class of REAPER-PRISM models: 2-bit and 4-bit lossless SigRoundV2 distill finetunes. Keep the support coming for upcoming drops!