rockypod/neotoi-coder

Neotoi Coder v1

A Rust/Dioxus 0.7 specialist fine-tuned from Qwen3-Coder-14B using RAFT (Retrieval-Augmented Fine-Tuning). Optimized for production-quality Dioxus 0.7 components with Tailwind v4 and WCAG 2.2 AAA accessibility.

Exam Results

Tier	Score	Required	Status
T1 Fundamentals	9/10	9/10	✅
T2 RSX Syntax	9/10	8/10	✅
T3 Signal Hygiene	10/10	8/10	✅
T4 WCAG/ARIA	9/10	7/10	✅
T5 use_resource	4/5	4/5	✅
T6 Hard Reasoning	2/5	2/5	✅
T7 Primitives+CSS	8/10	6/10	✅
Overall	51/60	50/60	✅ PASS

Model Details

• Base model: Qwen3-Coder-14B
• Method: RAFT (Retrieval-Augmented Fine-Tuning)
• Dataset: 3,156 curated Dioxus 0.7 examples
• Scope: Rust + Dioxus 0.7 + Tailwind v4 + WCAG 2.2 AAA
• Quantization: Q4_K_M (8.38 GB)
• Author: Kevin Miller, Jr.

Enabling Thinking Mode

This model supports Qwen3 native thinking tokens. Thinking must be enabled manually depending on your inference backend.

LM Studio

In the chat interface go to the prompt template settings and configure:

Field	Value
Before System	<|im_start|>system
After System	<|im_end|>
Before User	<|im_start|>user
After User	<|im_end|>
Before Assistant	<|im_start|>assistant\n<think>
After Assistant	<|im_end|>

Ollama

Create a Modelfile:

FROM neotoi-coder-v1-q4_k_m_final.gguf
PARAMETER temperature 0.2
PARAMETER num_predict 4096
PARAMETER repeat_penalty 1.15
PARAMETER stop "<|im_end|>"
TEMPLATE """<|im_start|>system
{{ .System }}<|im_end|>
<|im_start|>user
{{ .Prompt }}<|im_end|>
<|im_start|>assistant
<think>
"""
SYSTEM You are Neotoi, an expert Rust and Dioxus 0.7 developer. Always think step-by-step before answering.

llama.cpp / llama-cli

./llama-cli \
  -m neotoi-coder-v1-q4_k_m_final.gguf \
  -ngl 99 \
  --temp 0.2 \
  -p "<|im_start|>user\nYour question here<|im_end|>\n<|im_start|>assistant\n<think>"

What It Knows

• Dioxus 0.7 RSX brace syntax — never function-call style
• use_signal, use_resource with correct three-arm match
• r#for on label elements only, never inputs
• WCAG 2.2 AAA: aria_labelledby, aria_describedby, role="alert", role="dialog", live regions
• dioxus-primitives — no manual ARIA on managed components
• styles!() macro for CSS modules
• Tailwind v4 utility classes

What It Does Not Know

• Tier 6 hard reasoning edge cases (use_context panic behavior, optimistic UI race conditions) — known weak spots
• Playwright/E2E testing (out of scope)
• Non-Dioxus web frameworks

License

Neotoi Coder Community License v1.0 — see LICENSE file. Commercial use of model outputs permitted. Weight redistribution prohibited. Mental health deployment requires written permission.

Credits

Built with:

• Unsloth — 2x faster fine-tuning
• TRL — SFTTrainer
• Qwen3-Coder-14B — base model
• MLX — dataset generation on Apple Silicon
• Claude Code — dataset pipeline and training infrastructure
• Ansible — server automation and RAFT workflow orchestration
• repomix — bundling framework source into LLM context
• Forgejo — self-hosted git, source stored locally
• Zed — editor used throughout development
• Dioxus — the framework this model specializes in

Developed on:

• Apple M3 MacBook Pro — dataset generation, MLX inference, LM Studio
• Rocky Linux 10.1 — dataset generation, Unsloth fine-tuning, PyTorch, GGUF export
• CachyOS — additional RAFT pipeline work