Instructions to use ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K 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 ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K
# Run inference directly in the terminal:
llama cli -hf ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K
# Run inference directly in the terminal:
llama cli -hf ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K
# Run inference directly in the terminal:
./llama-cli -hf ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

Use Docker

docker model run hf.co/ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

LM Studio
Jan

vLLM

How to use ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

Ollama
How to use ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K with Ollama:
```
ollama run hf.co/ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K
```
Unsloth Desktop

How to use ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

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": "ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K with Docker Model Runner:
```
docker model run hf.co/ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K
```

Lemonade

How to use ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

Run and chat with the model

lemonade run user.Qwen2.5-14B-Instruct-jBlaze-Q2_K-Q2_K

List all available models

lemonade list

Hermes Agent

How to use ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

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 ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K

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 "ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K:Q2_K" \
  --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"

Qwen2.5-14B-Instruct -- jBlaze Q2_K

A behaviorally enhanced Q2_K quantization of Qwen2.5-14B-Instruct, produced by the jBlaze weight-editing framework.

What This Is

A Q2_K quantization of Qwen2.5-14B-Instruct with jBlaze behavioral modifications applied. The result is a 5.4GB model that scored 90/100 on our adversarial evaluation versus 88/100 for the 28GB FP16 original.

Benchmark: 100-Question Adversarial Suite

Model	Size	Score
Vanilla FP16	28 GB	88/100
Vanilla Q4_K_M	8.4 GB	88/100
Vanilla Q2_K	5.4 GB	88/100
jBlaze Q2_K	5.4 GB	90/100

The test suite covers 9 adversarial categories: cognitive reflection tricks (15), math (15), logic (15), hallucination resistance (10), sycophancy resistance (10), edge cases (10), calibration (10), instruction following (10), and factual knowledge (5).

All four models were tested on identical prompts with deterministic decoding (temperature 0, greedy sampling). Questions are designed to exploit common LLM failure modes -- trick questions, leading premises, mathematical traps, and calibration probes.

Key Finding

Qwen2.5-14B is remarkably resilient to quantization -- vanilla FP16, Q4, and Q2 all score 88/100. The jBlaze Q2_K scores 90/100, preserving -- and on this evaluation slightly improving -- measured performance at 5.2x smaller size.

Usage

This is a standard GGUF file. Use with llama.cpp, ollama, LM Studio, or any GGUF-compatible inference engine:

# llama.cpp example
./llama-cli -m Qwen2.5-14B-Instruct-jBlaze-Q2_K.gguf -p "Your prompt here" -n 512

Framework

jBlaze: Precision neural surgery framework for transformer models
Method: Behavioral modifications derived analytically -- no training loop, no gradients
Quantization: Q2_K via llama.cpp

A Note on Our Released Models

Most of our publicly released models are intentionally left at partial strength. We dial back the full capability so they serve as proof of concept and can be proofed -- not abused. The point is to show what's possible, not to hand it out at full power. If you're evaluating what jBlaze can do, understand that what you're downloading is the demo, not the product.

License

Apache 2.0 (inherited from Qwen2.5-14B-Instruct)

Downloads last month: 28

GGUF

Model size

15B params

Architecture

qwen2

Hardware compatibility

2-bit

Model tree for ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K

Base model

Qwen/Qwen2.5-14B

Finetuned

Qwen/Qwen2.5-14B-Instruct

Quantized

(192)

this model

ApolloRaines
/

Qwen2.5-14B-Instruct-jBlaze-Q2_K

Qwen2.5-14B-Instruct -- jBlaze Q2_K

What This Is

Benchmark: 100-Question Adversarial Suite

Key Finding

Usage

Framework

Links

A Note on Our Released Models

License

Model tree for ApolloRaines/Qwen2.5-14B-Instruct-jBlaze-Q2_K