macadeliccc

/

MonarchLake-7B

like 4

Yeah I can upload them for you tomorrow!

Thanks! What would be the prompt format to get the best output? Especially for chat use case?

https://huggingface.co./macadeliccc/MonarchLake-7B-GGUF/tree/main

The prompt template should be chatML. If youre using the GGUF version with ooba or llama.cpp it should auto populate from the config.

If chatml doesnt work try alpaca. sometimes rewriting the template doesnt work right if you merge lora weights instead of a full finetune.

Thanks! I'll try it in LM Studio with chatML. Is the context window 32k?

Yes it is 32k