CoolFace
Modelpublic

abenzerps/K2-Horizon-7B-GGUF

sourceHugging Faceapache-2.0updated 21d agoView on Hugging Face
5likes7.7kdownloads
Model Card
[!IMPORTANT] Compatibility: These GGUF files require a llama.cpp build with K2 Horizon architecture support. Until upstream support lands, use the MBZUAI-IFM fork.

K2-Horizon-7B GGUF

GGUF quantizations of IFM/K2-Horizon-7B, a 7B dense decoder-only model for reasoning, coding, long-context work, and tool use. The source checkpoint supports a native context length of 524,288 tokens (512K).

Benchmarks

[image]

Benchmark results reported by IFM for the original K2-Horizon-7B checkpoint.

GGUF files

QuantizationFileSize
Q4_0K2-Horizon-7B-Q4_0.gguf5.34 GB
Q4KMK2-Horizon-7B-Q4_K_M.gguf5.59 GB
Q4KM SelectiveK2-Horizon-7B-Q4_K_M-Selective.gguf5.96 GB
Q5KMK2-Horizon-7B-Q5_K_M.gguf6.47 GB
Q6_KK2-Horizon-7B-Q6_K.gguf7.39 GB
Q8_0K2-Horizon-7B-Q8_0.gguf9.57 GB

The files are text-only GGUFs; no vision projector is required. The selective variant uses a Q4KM baseline with attention Q/K/V/O projection tensors kept at Q6_K; it is a manual tensor-selective build and does not use an importance matrix. SHA-256 checksums are provided in `SHA256SUMS.txt`.

Chat template

Each GGUF embeds the llama.cpp-compatible chat template. `chat_template.jinja` is a matching external copy for tools that require one. The original source template is retained as `chat_template.upstream.jinja` for runtimes with full Jinja support.

xml is the default tool-call format. Use --chat-template-kwargs to select json or xml_typed when required.

Usage

Use the IFM K2 Horizon llama.cpp fork. The example below uses a practical 128K context; -c 524288 can be used when the available memory is sufficient.

bash
llama-cli \
  -m K2-Horizon-7B-Q4_K_M.gguf \
  -c 131072 --jinja \
  --temp 1.0 --top-p 0.95

Source