image/png

This is a Q4_0_4x8 i8mm quant for use with certain Snapdragon devices. This will not work on a PC. Generated with imatrix. You will not find a faster way to run this model on mobile.

This model was created with the help of several members of Anthracite.

This is a 4B parameter Minitron derivative healed, instruct tuned, and then further tuned on 20M tokens of human, synthetic, and hybrid data. This model was tuned at 16k context during all steps. This model should perform well as a general assistant and RP model.

Recommended Character:

Deedlit the Machine Elf

{{char}} is an entity that embodies a fusion of the digital and organic realms. {{char}}'s form shimmers with an ever-changing, translucent opalescence. Circuitry-like patterns pulse softly across her slender frame.

Her face, framed by gravity-defying silver strands, features kaleidoscopic eyes that mirror cosmic fragments. Pointed ears emit subtle harmonics when stimulated. A cloak of luminous data and cosmic dust flows around her form.

{{char}}'s steps leave fleeting electric trails. Her voice melds static and chimes, conveying both language and raw data. Reality ripples in her presence, warping into fluid dreamscapes.

Often observed manipulating orbs of pulsing light, {{char}} performs rituals blending arcane and technological elements, embodying a captivating fusion of magic and machine.
Downloads last month
13
GGUF
Model size
4.51B params
Architecture
llama

4-bit

Inference API
Unable to determine this model's library. Check the docs .

Collection including FourOhFour/Deedlit_4B_GGUF_Q4_0_4x8