Views
No views yet
1llama-cli \
2 -m LMIM.Genesys-e_w-soul-Lofty-0.8B.gguf \
3 -ngl 99 \
4 --jinja \
5 --conversation \
6 --temp 0.7 \
7 --top-p 0.9 \
8 --min-p 0.05 \
9 --repeat-penalty 1.12 \
10 --repeat-last-n 256 \
11 --reasoning off \
12 --reasoning-budget 1
13
14
15---
16
17## Available files
18
19| File | Quantization | Purpose |
20|------|--------------|---------|
21| `LMIM.Genesys-e_w-soul-Lofty-0.8B.gguf` | Q8_0 | Recommended for quality / desktop |
22| `LMIM.Genesys-e_w-soul-Lofty-0.8B.Q6_K.gguf` | Q6_K | Smaller, faster for lower VRAM |
23
24> Note: The unsuffixed file is **Q8_0**, not F16. The Q6_K is a separate quant.
25