K2 Horizon 3.7B MLX 8-bit

MLX conversion of IFM/K2-Horizon-3.7B, created specifically for MLXHub with the DreamFoundries mlx-lm fork at 0f74c0e. Affine 8-bit quantization uses group size 64. The K2 routers (mlp.gate and, where present, self_attn.v_router) remain unquantized by the model implementation.

The original model is by Institute of Foundation Models (IFM) and is released under Apache-2.0. Comparative quality and performance benchmarks are not available for this conversion.

Use with MLX

from mlx_lm import load, generate
model, tokenizer = load("DreamFoundries/K2-Horizon-3.7B-MLX-8bit")
response = generate(model, tokenizer, prompt="hello", verbose=True)

---