GGUF models by Kaoeiri
70 GGUF models quantized by Kaoeiri on Hugging Face. Sizes run from 7.25B to 39B parameters. The longest trained context is 128K tokens.
About Kaoeiri on Hugging Face
Kaoeiri has published 70 GGUF models that ggufmodels.com tracks, across families such as Magpantheonselrp, Magpantheonsel and Keiana. The most downloaded are Keiana-L3-Test5.75-8B-13.5, Hedone L3.1 8B v1.1 Q8 0 and Mistral magnum X Cydonia v1.3 M Q4 K M.
The quants Kaoeiri publishes most often are Q4_K_M (46), Q6_K (8), Q3_K_M (6) and Q8_0 (4). Their Q4_K_M files range from 13.3 to 19.8 GB. Parameter counts run from 7.25B to 39B, and the longest trained context window is 128K tokens.
Frequently asked questions
What GGUF models has Kaoeiri published?
70 models are listed here, mainly from the Magpantheonselrp, Magpantheonsel and Keiana families. The most downloaded is Keiana-L3-Test5.75-8B-13.5.
Which quants does Kaoeiri publish?
The most common are Q4_K_M (46 models), Q6_K (8 models), Q3_K_M (6 models) and Q8_0 (4 models).