hako-koi

Mid tier (7B) of the hako local model family from Mithraeum. A stock wrap of Qwen2.5-Coder-7B-Instruct, converted to the MLF2 container and run by the from-scratch hako engine (hakm) — no llama.cpp, no ggml, no ollama at runtime.

What this is

Base: stock Qwen2.5-Coder-7B-Instruct — no weight fine-tune yet. A real fine-tune earns a versioned id (hako-koi-v0.0.1).
Format: MLF2 (k-quant blocks copied verbatim from the source GGUF), mmap, int8-activation × int4-weight fast path.
Template: ChatML (<|im_start|> / <|im_end|>).

Use it

hako                       # then  :pull hako-koi
# or:
hakm hako-koi.mlf2 --sys "You are hako." "write a B-tree in C"

Resolves to https://huggingface.co/mithraeum/hako-koi/resolve/main/hako-koi.mlf2.

License

Apache-2.0 (the 7B base is Apache-licensed). See LICENSE and NOTICE. The hako engine code is GPL-3.0; these weights are Apache-2.0.

Downloads last month: -; Downloads are not tracked for this model. How to track

Model tree for mithraeum/hako-koi

Base model

Qwen/Qwen2.5-7B

Finetuned

Qwen/Qwen2.5-Coder-7B

Finetuned

Qwen/Qwen2.5-Coder-7B-Instruct

Finetuned

(393)

this model