Sft GGUF models
115 GGUF models in the Sft family, quantized by 17 uploaders. Sizes run from 21M to 32.8B parameters. The longest trained context is 1000K tokens.
About the Sft GGUF models
Sft is represented on ggufmodels.com by 115 GGUF models, quantized by 17 uploaders on Hugging Face. The most downloaded are MT7Bi sft i1, 0810 SFT Github Orlm Industryor 14B 5e 6 64 and Sft_Subtitles_Fixer_V1_Merged. Parameter counts run from 21M to 32.8B, and the longest trained context window is 1000K tokens.
A Q4_K_M file of a Sft model is between 0.1 to 19.9 GB, which is the quant most people start with. Most files come from mradermacher (32), ApocalypseParty (23) and Andrei1980 (16). The quants you will meet most often are Q8_0 (63), Q4_K_M (60), Q2_K (60) and Q6_K (54), where the number is how many models offer that quant.
Frequently asked questions
Which Sft GGUF should I download?
The most downloaded Sft GGUF here is MT7Bi sft i1. 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 Sft model need?
Q4_K_M files of Sft models on this site range from 0.1 to 19.9 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 Sft GGUF files come from?
The most common quantizers here are mradermacher, ApocalypseParty and Andrei1980. They convert the original model and publish quantized files on Hugging Face. ggufmodels.com does not host files; every download button opens Hugging Face.