Instructions to use siamcafe/bom-ai with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use siamcafe/bom-ai with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf siamcafe/bom-ai:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf siamcafe/bom-ai:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf siamcafe/bom-ai:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf siamcafe/bom-ai:Q4_K_M

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf siamcafe/bom-ai:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf siamcafe/bom-ai:Q4_K_M

Use Docker

docker model run hf.co/siamcafe/bom-ai:Q4_K_M

Llama new
LM Studio
Jan
Ollama
How to use siamcafe/bom-ai with Ollama:
```
ollama run hf.co/siamcafe/bom-ai:Q4_K_M
```
Unsloth Desktop

How to use siamcafe/bom-ai with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf siamcafe/bom-ai:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "siamcafe/bom-ai:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use siamcafe/bom-ai with Docker Model Runner:
```
docker model run hf.co/siamcafe/bom-ai:Q4_K_M
```

Lemonade

How to use siamcafe/bom-ai with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull siamcafe/bom-ai:Q4_K_M

Run and chat with the model

lemonade run user.bom-ai-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use siamcafe/bom-ai with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf siamcafe/bom-ai:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default siamcafe/bom-ai:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use siamcafe/bom-ai with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf siamcafe/bom-ai:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "siamcafe/bom-ai:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

BOM AI v0.3.2

ผู้ช่วย AI ภาษาไทย ที่เห็นภาพและอ่านภาษาไทยได้ — ฟรี ใช้งานแบบออฟไลน์

Thai-first Vision-Language AI model. Official stable release, 2026-10-30. Promoted from v0.3.1-rc.1 after all 6 ship gates passed.

What's New in v0.3.2 (2026-10-30)

🎯 Official Release — Promoted from v0.3.1-rc.1 to stable v0.3.2
✅ All 6 Ship Gates PASSED — Vision, Identity, Zero-Han (250/250=0), Thai≥9.0, Refusal, Regression
🔒 Zero-Han Solution — GBNF grammar allowlist eliminates all CJK character leakage
👁️ Vision (VL) Verified — Qwen3.5-4B multimodal inference confirmed working
🤖 Identity Correct — Model responds as "BOM" with proper Thai identity
⚙️ GOD CONFIG v7.0 — Production inference parameters locked and deployed fleet-wide

Model Specs

Property	Value
Base Model	DavidAU/Qwen3.5-4B-Claude-4.6-HighIQ-THINKING
Architecture	Qwen3-VL (Vision-Language, 4B params)
Quantization	Q4_K_M GGUF (~3.4 GB)
Training	3 rounds SFT + repair (PhaseA 465 steps, repair_r2 157, repair_r3 162)
License	Apache 2.0
Zero-Han	GBNF grammar allowlist — 0 CJK leakage

Files & Checksums (SHA256)

e2e460e979a24bba9187854b96e79fbab5561b26c2493ffcee00bf659212fd67  model-Q4_K_M.gguf
ea8eb56d7fdcaa49698f1a06c61d90ab18c30ac96f61f02054a0f36c0bc3c2b9  mmproj-f16.gguf
b1b9b980cb6e9cb0cb0c176e68e379496d65f3265f0141c00103ac02943f4875  model-bf16.gguf

Weights are identical to v0.3.1-rc.1 — this release is a quality-gate promotion, no retraining.

How to Use

Ollama

ollama pull siamcafe/bom-ai:latest
ollama run siamcafe/bom-ai:latest

llama.cpp / ComfyUI / any GGUF runtime

huggingface-cli download siamcafe/bom-ai

Load model-Q4_K_M.gguf with mmproj-f16.gguf as the vision projector.

Ship Gates (ALL PASS ✅)

G1 Vision ✅ — VL model verified
G2 Identity ✅ — "ผมชื่อ BOM" correct
G3 Zero-Han ✅ — 250/250 = 0 Han characters
G4 Thai≥9.0 ✅ — Quality verified
G5 Refusal ✅ — Safety responses verified
G6 Regression ✅ — No regression in mega tests

Changelog

v0.3.2 (2026-10-30)

Official stable release. All 6/6 ship gates passed. Zero-Han GBNF verified. VL vision working. GOD CONFIG v7.0 deployed.

v0.3.1-rc.1 (2026-10-28)

First release candidate. Q4_K_M ORPO + GBNF export.

Downloads last month: 461

GGUF

Model size

4B params

Architecture

qwen35

Hardware compatibility

4-bit

16-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for siamcafe/bom-ai

Base model

Qwen/Qwen3.5-4B-Base

Finetuned

Qwen/Qwen3.5-4B

Finetuned

DavidAU/Qwen3.5-4B-Claude-4.6-HighIQ-THINKING

Quantized

(4)

this model