Qwopus3.6-27B-Coder-MTP-BF16.gguf to i1

#2635
by Dumbledore2 - opened

Jackrong/Qwopus3.6-27B-Coder-MTP-GGUF , Qwopus3.6-27B-Coder-MTP-BF16.gguf

to
i1 - IQ2 , IQ3 , IQ4

ty

https://huggingface.co/Jackrong/Qwopus3.6-27B-Coder-MTP-GGUF/blob/main/Qwopus3.6-27B-Coder-MTP-BF16.gguf

can I have safetensors please ?

Dumbledore2 changed discussion status to closed

https://huggingface.co/mradermacher/Qwopus3.6-35B-A3B-Coder-i1-GGUF/tree/main

Im confused, if exists why request lol
mtp ? it is usually badly supported by the llama cpp, it might need patching.
Does mtp exist in our quants that is listen? im not sure, and cannot test right now

https://huggingface.co/Jackrong/Qwopus3.6-35B-A3B-Coder-FP8
is fp8 good ?

what's the point when we have a full weights quantized already? it will either be on same level or worse...

https://huggingface.co/mradermacher/Qwopus3.6-35B-A3B-Coder-i1-GGUF/tree/main

Im confused, if exists why request lol
mtp ? it is usually badly supported by the llama cpp, it might need patching.
Does mtp exist in our quants that is listen? im not sure, and cannot test right now

https://huggingface.co/Jackrong/Qwopus3.6-35B-A3B-Coder-FP8
is fp8 good ?

what's the point when we have a full weights quantized already? it will either be on same level or worse...

--- Im confused, if exists why request lol
miss label non mtp
----- Does mtp exist in our quants that is listen? im not sure, and cannot test right now
yes i checked using llama-server with mtp on so its also miss labeled
----- what's the point when we have a full weights quantized already? it will either be on same level or worse...
cos miss labeled i didnt knew there tensors files with mtp

Sign up or log in to comment