GGUF models by MilyaShams
49 GGUF models quantized by MilyaShams on Hugging Face. Sizes run from 1.72B to 7.61B parameters. The longest trained context is 40K tokens.
About MilyaShams on Hugging Face
MilyaShams has published 49 GGUF models that ggufmodels.com tracks, across families such as Qwen and Lite. The most downloaded are Qwen3 1.7B SparseGPT unstruct 0.3 bs128 damp0.01, Qwen3 1.7B SmoothQuant 0.4 AutoRound ign and Qwen3 1.7B Wanda unstruct 0.1.
The quants MilyaShams publishes most often are Q8_0 (22), Q4_K_M (18) and F16 (9). Their Q4_K_M files range from 1.1 to 4.7 GB. Parameter counts run from 1.72B to 7.61B, and the longest trained context window is 40K tokens.
Frequently asked questions
What GGUF models has MilyaShams published?
49 models are listed here, mainly from the Qwen and Lite families. The most downloaded is Qwen3 1.7B SparseGPT unstruct 0.3 bs128 damp0.01.
Which quants does MilyaShams publish?
The most common are Q8_0 (22 models), Q4_K_M (18 models) and F16 (9 models).