MiniMax onX: Weights dropping REALLY, REALLY, SOON Posted by XMasterrrr@reddit | LocalLLaMA | View on Reddit | 10 comments
[-] FullstackSensei@reddit IIRC, Ahmed was a regular on this sub Reply Submit [-] No_Afternoon_4260@reddit He is, look for the ultimate rig Reply Submit
[-] rm-rf-rm@reddit Weights are released, continue discussion in release thread: https://www.reddit.com/r/LocalLLaMA/comments/1r3pxy7/minimaxaiminimaxm25_hugging_face/ Reply Submit
[-] Potential_Block4598@reddit When is the REAP, REAM or THRIFT version of that dropping ?! Along with idk Q2ks Intel autoround or INT4 Or MXFP4 Canβt wait please ππ» Reply Submit [-] No_Afternoon_4260@reddit I only use fp32 even for kimi /s Reply Submit [-] Potential_Block4598@reddit Wow You mean you have like 4 Terabytes of RAM maybe idk ?! (If fully GPU that means you have like $20M 64 GPU cluster at your garage maybe ?!) And idk how much active parameters is that ?! 32b that is idk 128GB of VRAM Nice bro nice Reply Submit [-] No_Afternoon_4260@reddit I only run inference straight up from a zfs pool on spinning rust /s π Reply Submit [-] Potential_Block4598@reddit What ?? Reply Submit [-] Potential_Block4598@reddit I though that could work somehow Reply Submit [-] No_Afternoon_4260@reddit Don't you get the "/s"? This is for sarcasm Reply Submit
[-] No_Afternoon_4260@reddit I only use fp32 even for kimi /s Reply Submit [-] Potential_Block4598@reddit Wow You mean you have like 4 Terabytes of RAM maybe idk ?! (If fully GPU that means you have like $20M 64 GPU cluster at your garage maybe ?!) And idk how much active parameters is that ?! 32b that is idk 128GB of VRAM Nice bro nice Reply Submit [-] No_Afternoon_4260@reddit I only run inference straight up from a zfs pool on spinning rust /s π Reply Submit [-] Potential_Block4598@reddit What ?? Reply Submit [-] Potential_Block4598@reddit I though that could work somehow Reply Submit [-] No_Afternoon_4260@reddit Don't you get the "/s"? This is for sarcasm Reply Submit
[-] Potential_Block4598@reddit Wow You mean you have like 4 Terabytes of RAM maybe idk ?! (If fully GPU that means you have like $20M 64 GPU cluster at your garage maybe ?!) And idk how much active parameters is that ?! 32b that is idk 128GB of VRAM Nice bro nice Reply Submit [-] No_Afternoon_4260@reddit I only run inference straight up from a zfs pool on spinning rust /s π Reply Submit [-] Potential_Block4598@reddit What ?? Reply Submit [-] Potential_Block4598@reddit I though that could work somehow Reply Submit [-] No_Afternoon_4260@reddit Don't you get the "/s"? This is for sarcasm Reply Submit
[-] No_Afternoon_4260@reddit I only run inference straight up from a zfs pool on spinning rust /s π Reply Submit [-] Potential_Block4598@reddit What ?? Reply Submit [-] Potential_Block4598@reddit I though that could work somehow Reply Submit [-] No_Afternoon_4260@reddit Don't you get the "/s"? This is for sarcasm Reply Submit
[-] Potential_Block4598@reddit What ?? Reply Submit [-] Potential_Block4598@reddit I though that could work somehow Reply Submit [-] No_Afternoon_4260@reddit Don't you get the "/s"? This is for sarcasm Reply Submit
10 Comments
FullstackSensei@reddit
No_Afternoon_4260@reddit
rm-rf-rm@reddit
Potential_Block4598@reddit
No_Afternoon_4260@reddit
Potential_Block4598@reddit
No_Afternoon_4260@reddit
Potential_Block4598@reddit
Potential_Block4598@reddit
No_Afternoon_4260@reddit