README / README.md
RhinoWithAcape's picture
Reposition: the small-GPU and modest-RAM champion, for everyone
f24cfd6 verified
|
Raw
History Blame Contribute Delete
1.62 kB
---
title: README
emoji: πŸ”₯
colorFrom: yellow
colorTo: red
sdk: static
pinned: false
---
# Hob Forge
**Small household spirits for small machines.**
A hob is an English household spirit that quietly does the work. Hob Forge exists for one
reason: **to help you actually run good models on the hardware you already have.** We build
**3B–4B finetunes and merges β€” the best little agentic models** β€” and we are unapologetically
**the champion of small GPUs and modest RAM**: every release states exactly what it needs at
8GB, 12GB, and 16GB (and what CPU-only costs you), and ships with the exact way to run it β€”
step-by-step guides, tested configs, real support.
Every release promises:
- **A stated memory budget.** Quant sizes and KV-cache maths per hardware tier β€” 8GB, 12GB,
16GB β€” in every card, measured, not projected. If it doesn't fit with real context at a
tier, the card says so plainly.
- **Honest evals.** Held-out benchmarks only, decontamination stated, baselines included β€”
the raw base and a same-size peer, same harness, same settings. We publish the harness
output, not vibes.
- **Full provenance.** Which teachers, which datasets, which licenses, what filtering β€”
disclosure tables with verified-trace counts.
- **A job, not a persona.** Each hob does one thing well β€” terminal work, code fixing,
tool routing. Merges combine proven specialists, and a merge that doesn't beat its
parents doesn't get a name.
First resident in the works: **smeagle** β€” a 4B terminal/software-engineering agent.
*Built in a home lab in the UK on a strict diet of verified data.*