---
license: apache-2.0
tags:
- gguf
- fusewalk
base_model:
- Qwen/Qwen3-0.6B
- Qwen/Qwen3-1.7B
- Qwen/Qwen3.5-4B
---

# Fusewalk models

Fine-tuned models for Fusewalk, a crafting game: give a model two items and it answers what they make.

Each model is a 4-bit (Q4_K_M) GGUF for llama.cpp. [`index.json`](index.json) lists them with the exact prompt template and stop strings each one needs.

## Modifications

Each model is a fine-tune of the base model listed below. It was trained (LoRA merged into the weights, or a full fine-tune) on the Fusewalk Codex, a hand-written set of item-combination recipes, so that it answers with the names of the items two items make. It was then converted to GGUF and quantised to 4-bit. The base models' licences and notices are in [`licenses/`](https://huggingface.co/npateman/fusewalk-models/tree/main/licenses), copied unchanged from the base model repos.

## Models

| Model | Fine-tune of | Training | Licence | Codex | Size | Plausible |
|---|---|---|---|---|---|---|
| [20261016-152956-qwen3-1.7b-v0.6](https://huggingface.co/npateman/fusewalk-models/resolve/main/20261016-152956-qwen3-1.7b-v0.6.gguf) | [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B) | LoRA | [apache-2.0](https://huggingface.co/npateman/fusewalk-models/blob/main/licenses/Qwen--Qwen3-1.7B/LICENSE) | v0.6 | 1,056 MB | 51% |
| [20261015-090302-qwen3-1.7b-v0.4](https://huggingface.co/npateman/fusewalk-models/resolve/main/20261015-090302-qwen3-1.7b-v0.4.gguf) | [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B) | LoRA | [apache-2.0](https://huggingface.co/npateman/fusewalk-models/blob/main/licenses/Qwen--Qwen3-1.7B/LICENSE) | v0.4 | 1,056 MB | 57% |
| [20261015-090921-qwen3-0.6b-v0.4](https://huggingface.co/npateman/fusewalk-models/resolve/main/20261015-090921-qwen3-0.6b-v0.4.gguf) | [Qwen/Qwen3-0.6B](https://huggingface.co/Qwen/Qwen3-0.6B) | full fine-tune | [apache-2.0](https://huggingface.co/npateman/fusewalk-models/blob/main/licenses/Qwen--Qwen3-0.6B/LICENSE) | v0.4 | 378 MB | 41% |
| [20261014-135949-qwen3.5-4b-v0.4](https://huggingface.co/npateman/fusewalk-models/resolve/main/20261014-135949-qwen3.5-4b-v0.4.gguf) | [Qwen/Qwen3.5-4B](https://huggingface.co/Qwen/Qwen3.5-4B) | LoRA | [apache-2.0](https://huggingface.co/npateman/fusewalk-models/blob/main/licenses/Qwen--Qwen3.5-4B/LICENSE) | v0.4 | 2,583 MB | 74% |

**Plausible** is the share of answers that make sense, on pairs held out from training: an LLM judge rates each answer good, mash-up (the two names glued together), weak or nonsense, and this counts the good ones. There's often more than one sensible answer, so it says more than an exact match with the Codex would.

## Try it

The prompt is two quoted item names, in alphabetical order, joined with ` + `, inside the model's prompt template. The answer is one or two quoted names on the first line. With [llama.cpp](https://github.com/ggml-org/llama.cpp) and the newest model:

```sh
llama-completion -m 20261016-152956-qwen3-1.7b-v0.6.gguf -no-cnv --temp 0 -n 24 \
  -p '<|im_start|>user\n"Fire" + "Water"<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\n'
```

It answers something like:

```
"Steam"
```

Each model's template is `prompt_template` in [`index.json`](index.json), with `{pair}` where the items go. Stop at the first newline.