Modelpublic

TribeBlend/tribeblend-etl-kimi-linear-48b-a3b-instruct

sourceHugging Faceotherupdated 3mo agoView on Hugging Face

0likes15downloads

Model Card

tribeblend-etl-kimi-linear-48b-a3b-instruct

Kimi Linear 48B A3B Instruct fine-tuned for high-capability local systems without Moonshot APIs.

Direct base-model GGUF (Q4KM) of moonshotai/Kimi-Linear-48B-A3B-Instruct, published for TribeBlend's local Data Chat runtime. TribeBlend grounds answers with Knowledge Graph context at prompt time and a model-aware agent harness, so the base instruction/reasoning model ships as-is (no fine-tuning).

—Base model: moonshotai/Kimi-Linear-48B-A3B-Instruct
—Provider / family: kimi / kimi-linear
—Local runtime arch: kimi-linear
—Recommended profile: expert
—Quantization: Q4KM
—Native context window: 1048576

Usage

Designed for TribeBlend Data Chat, loaded via llama-cpp-2.

License

Inherits the upstream base-model license (other); verify upstream terms before redistribution.