GEHIRNS/magi-1.1-2B
Text Generation • 2B • Updated • 745 • 1
The magi-1.0 recipe plus a LoRA that stops loops and finishes thinking. magi-1.1-4B: 81.7%, density 0.939. magi-1.1-2B: density 1.149.
Note Qwen3.5-2B with a think-less LoRA and the magi-1.0 recipe. 0.83 GiB GGUF, density 1.149 at 32K with a q4_0 KV cache (magi-1.0: 0.903 at the same settings), 338 tok/s on an RTX 5070.