ExpedientFalcon/
Qwen3-4B-UD-Q5_K_XL:latest

91 5 months ago

Qwen3-4B Q5_K_XL Unsloth UD 2.0 adaptively quantized model, much better for coding than vanilla Q4_K_M quants without taking up the VWAM of an 8-bit Q8_0 model. From https://huggingface.co/unsloth/Qwen3-4B-GGUF/tree/main

tools
... /
params
b00592e9d0ee · 166B
{
"min_p": 0,
"num_ctx": 40960,
"num_predict": 16384,
"repeat_penalty": 1,
"stop": [
"<|im_start|>",
"<|im_end|>"
],
"temperature": 0.6,
"top_k": 20,
"top_p": 0.95
}