Lite GGUF models
92 GGUF models in the Lite family, quantized by 39 uploaders. Sizes run from 64M to 22.2B parameters. The longest trained context is 256K tokens.
About the Lite GGUF models
Lite is represented on ggufmodels.com by 92 GGUF models, quantized by 39 uploaders on Hugging Face. The most downloaded are T lite it 2.1, vllm-d3-lite and T Lite It 1.0. Parameter counts run from 64M to 22.2B, and the longest trained context window is 256K tokens.
A Q4_K_M file of a Lite model is between 0.0 to 13.3 GB, which is the quant most people start with. Most files come from mradermacher (18), QuantFactory (7) and NikolayKozloff (6). The quants you will meet most often are Q8_0 (56), Q4_K_M (56), Q5_K_M (53) and Q2_K (49), where the number is how many models offer that quant.
Frequently asked questions
Which Lite GGUF should I download?
The most downloaded Lite GGUF here is T lite it 2.1. 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 Lite model need?
Q4_K_M files of Lite models on this site range from 0.0 to 13.3 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 Lite GGUF files come from?
The most common quantizers here are mradermacher, QuantFactory and NikolayKozloff. They convert the original model and publish quantized files on Hugging Face. ggufmodels.com does not host files; every download button opens Hugging Face.