GGUF models by DevQuasar-2
37 GGUF models quantized by DevQuasar-2 on Hugging Face. Sizes run from 135M to 123B parameters. The longest trained context is 986K tokens.
About DevQuasar-2 on Hugging Face
DevQuasar-2 has published 37 GGUF models that ggufmodels.com tracks, across families such as Qwen, Raptor and Phi. The most downloaded are Qwen.Qwen2 VL 7B Instruct, Granite 8b Code Base 128k and Granite 20b Code Instruct 8k.
The quants DevQuasar-2 publishes most often are Q5_K_M (36), Q4_K_M (36), Q3_K_M (36) and Q2_K (36). Their Q4_K_M files range from 0.1 to 73.2 GB. Parameter counts run from 135M to 123B, and the longest trained context window is 986K tokens.
Frequently asked questions
What GGUF models has DevQuasar-2 published?
37 models are listed here, mainly from the Qwen, Raptor and Phi families. The most downloaded is Qwen.Qwen2 VL 7B Instruct.
Which quants does DevQuasar-2 publish?
The most common are Q5_K_M (36 models), Q4_K_M (36 models), Q3_K_M (36 models) and Q2_K (36 models).