Kosolar GGUF models
73 GGUF models in the Kosolar family, quantized by 6 uploaders. Sizes run from 7.32B to 10.9B parameters. The longest trained context is 4K tokens.
About the Kosolar GGUF models
Kosolar is represented on ggufmodels.com by 73 GGUF models, quantized by 6 uploaders on Hugging Face. The most downloaded are KoSOLAR 10.9B v1.0, KoSOLAR 10.7B merge v3.4 i1 and KoSOLAR 10.9B v0.5 i1. Parameter counts run from 7.32B to 10.9B, and the longest trained context window is 4K tokens.
A Q4_K_M file of a Kosolar model is between 4.4 to 7.2 GB, which is the quant most people start with. Most files come from mradermacher (34), tensorblock (21) and RichardErkhov (14). The quants you will meet most often are Q2_K (71), IQ4_XS (51), Q5_K_M (50) and Q4_K_S (50), where the number is how many models offer that quant.
Frequently asked questions
Which Kosolar GGUF should I download?
The most downloaded Kosolar GGUF here is KoSOLAR 10.9B v1.0. Pick the size that fits your GPU first, then choose a quant: Q4_K_M is the usual starting point, and Q5_K_M or Q6_K give more quality if you have memory to spare. The VRAM pages below list what fits each GPU.
How much VRAM does a Kosolar model need?
Q4_K_M files of Kosolar models on this site range from 4.4 to 7.2 GB. Add the KV cache for your context length and about 0.5 GB for the runtime, which usually means 1.5 to 3 GB on top of the file at an 8K context.
Where do Kosolar GGUF files come from?
The most common quantizers here are mradermacher, tensorblock and RichardErkhov. They convert the original model and publish quantized files on Hugging Face. ggufmodels.com does not host files; every download button opens Hugging Face.