Instructions to use joebos/Second-1.0-2B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use joebos/Second-1.0-2B-GGUF 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 joebos/Second-1.0-2B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf joebos/Second-1.0-2B-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf joebos/Second-1.0-2B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf joebos/Second-1.0-2B-GGUF: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 joebos/Second-1.0-2B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf joebos/Second-1.0-2B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/joebos/Second-1.0-2B-GGUF:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use joebos/Second-1.0-2B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "joebos/Second-1.0-2B-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "joebos/Second-1.0-2B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/joebos/Second-1.0-2B-GGUF:Q4_K_M

Ollama
How to use joebos/Second-1.0-2B-GGUF with Ollama:
```
ollama run hf.co/joebos/Second-1.0-2B-GGUF:Q4_K_M
```
Unsloth Desktop

How to use joebos/Second-1.0-2B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf joebos/Second-1.0-2B-GGUF: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": "joebos/Second-1.0-2B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use joebos/Second-1.0-2B-GGUF with Docker Model Runner:
```
docker model run hf.co/joebos/Second-1.0-2B-GGUF:Q4_K_M
```

Lemonade

How to use joebos/Second-1.0-2B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull joebos/Second-1.0-2B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Second-1.0-2B-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use joebos/Second-1.0-2B-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf joebos/Second-1.0-2B-GGUF: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 joebos/Second-1.0-2B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use joebos/Second-1.0-2B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf joebos/Second-1.0-2B-GGUF: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 "joebos/Second-1.0-2B-GGUF: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"

Second-1.0-2B-GGUF: In-Browser Quantized Chess Model

This repository provides Q4_K_M GGUF weights for Second-1.0-2B, optimized for client-side in-browser execution with wllama (WebAssembly).

Base Model: second-chess/Second-1.0-2B
Repository: https://github.com/kleinebossie/second
Runtime Decision: docs/decisions/runtime.md

Model Files and Browser Shards

Browsers and Chrome extensions impose memory limits when allocating single large contiguous buffers. To support browser caching and streaming downloads, this repository provides both a single monolithic GGUF file and < 500 MB shards:

File Name	Size	Purpose
`Second-1.0-2B-Q4_K_M.gguf`	~1.22 GB	Standard monolithic file for `llama.cpp` CLI, Ollama, and desktop runtimes.
`Second-1.0-2B-Q4_K_M-00001-of-00003.gguf`	480 MB	Browser shard 1 of 3 (for extension side panels and IndexedDB caching).
`Second-1.0-2B-Q4_K_M-00002-of-00003.gguf`	480 MB	Browser shard 2 of 3.
`Second-1.0-2B-Q4_K_M-00003-of-00003.gguf`	261 MB	Browser shard 3 of 3.

Browser Benchmark Results (`wllama`)

Tested in Chrome Manifest V3 Side Panel on Intel Core Ultra 7 CPU:

Runtime	Quantization	Download Size	Time to First Token	Speed	Memory / Buffer	Status
wllama (Wasm)	`Q4_K_M`	1,221.5 MB	16,571 ms	14.38 tok/s	1024 MB Wasm Heap	Verified
wllama 0.8B fallback	`Q4_K_M`	507.8 MB	6,647 ms	28.11 tok/s	1024 MB Wasm Heap	Available

Note: WebLLM and Transformers.js failed shader compilation on Qwen 3.5's hybrid Gated DeltaNet attention kernels. wllama using WebAssembly operates reliably across all operating systems without GPU shader failures.

Post-Quantization Quality Check

Evaluated across 100 test prompts from eval/testset-1.0.jsonl with greedy decoding (temperature = 0):

Baseline Hard-Fail Rate (bfloat16): 4.67% (frozen 300 test items)
Quantized Hard-Fail Rate (Q4_K_M): ≤ 6.70% (PASSED release gate)
Target Length Compliance: 99.67% in-band
Format Pass Rate: 100.0%
First-Person Diary Voice: 0.00%

Usage in JavaScript / TypeScript with `wllama`

import { Wllama } from '@wllama/wllama';

const CONFIG_PATHS = {
  'single-thread/wllama.wasm': '/wllama/wllama-single.wasm',
  'multi-thread/wllama.wasm': '/wllama/wllama-multi.wasm',
};

const wllama = new Wllama(CONFIG_PATHS);

// Load sharded model files directly into browser cache
const shardUrls = [
  'https://huggingface.co/second-chess/Second-1.0-2B-GGUF/resolve/main/Second-1.0-2B-Q4_K_M-00001-of-00003.gguf',
  'https://huggingface.co/second-chess/Second-1.0-2B-GGUF/resolve/main/Second-1.0-2B-Q4_K_M-00002-of-00003.gguf',
  'https://huggingface.co/second-chess/Second-1.0-2B-GGUF/resolve/main/Second-1.0-2B-Q4_K_M-00003-of-00003.gguf',
];

await wllama.loadModelFromUrl(shardUrls, {
  n_ctx: 1024,
  n_threads: navigator.hardwareConcurrency || 4,
});

// Run greedy inference with Second chat template
const output = await wllama.createCompletion({
  prompt: promptText,
  temperature: 0.0,
  max_tokens: 200,
});
console.log(output);

Usage with `llama.cpp` CLI

llama-cli \
  -m Second-1.0-2B-Q4_K_M.gguf \
  -p "<|im_start|>system\nYou are Second, a chess assistant.<|im_end|>\n<|im_start|>user\ntask: explain_move\n...<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n" \
  -c 1024 \
  --temp 0

Fair Play Policy

Second is strictly for post-game chess review and study. It fails closed and refuses to operate during active live games.

License

Apache License 2.0.

Downloads last month: 128

GGUF

Model size

2B params

Architecture

qwen35

Hardware compatibility

4-bit

Second-1.0-2B-GGUF: In-Browser Quantized Chess Model

Model Files and Browser Shards

Browser Benchmark Results (wllama)

Post-Quantization Quality Check

Usage in JavaScript / TypeScript with wllama

Usage with llama.cpp CLI

Fair Play Policy

License

Browser Benchmark Results (`wllama`)

Usage in JavaScript / TypeScript with `wllama`

Usage with `llama.cpp` CLI