GGUF models by macadeliccc
42 GGUF models quantized by macadeliccc on Hugging Face. Sizes run from 3.21B to 69B parameters. The longest trained context is 195K tokens.
About macadeliccc on Hugging Face
macadeliccc has published 42 GGUF models that ggufmodels.com tracks, across families such as Llama, Mistral and Solar. The most downloaded are laser dolphin mixtral 2x7b dpo, WestLake 7B v2 laser truthy dpo and magistrate 3.2 3b it.
The quants macadeliccc publishes most often are Q5_K_M (38), Q4_K_M (37), Q8_0 (33) and Q2_K (33). Their Q4_K_M files range from 2.0 to 41.4 GB. Parameter counts run from 3.21B to 69B, and the longest trained context window is 195K tokens.
Frequently asked questions
What GGUF models has macadeliccc published?
42 models are listed here, mainly from the Llama, Mistral and Solar families. The most downloaded is laser dolphin mixtral 2x7b dpo.
Which quants does macadeliccc publish?
The most common are Q5_K_M (38 models), Q4_K_M (37 models), Q8_0 (33 models) and Q2_K (33 models).