Instructions to use MrFuzzihead/Nex-N2.5-mini-APEX-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 MrFuzzihead/Nex-N2.5-mini-APEX-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 MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16
# Run inference directly in the terminal:
llama cli -hf MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16
# Run inference directly in the terminal:
llama cli -hf MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

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 MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16
# Run inference directly in the terminal:
./llama-cli -hf MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

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 MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

Use Docker

docker model run hf.co/MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

LM Studio
Jan

vLLM

How to use MrFuzzihead/Nex-N2.5-mini-APEX-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MrFuzzihead/Nex-N2.5-mini-APEX-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": "MrFuzzihead/Nex-N2.5-mini-APEX-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

Ollama
How to use MrFuzzihead/Nex-N2.5-mini-APEX-GGUF with Ollama:
```
ollama run hf.co/MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16
```
Unsloth Desktop

How to use MrFuzzihead/Nex-N2.5-mini-APEX-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

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": "MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use MrFuzzihead/Nex-N2.5-mini-APEX-GGUF with Docker Model Runner:
```
docker model run hf.co/MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16
```

Lemonade

How to use MrFuzzihead/Nex-N2.5-mini-APEX-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

Run and chat with the model

lemonade run user.Nex-N2.5-mini-APEX-GGUF-BF16

List all available models

lemonade list

Hermes Agent

How to use MrFuzzihead/Nex-N2.5-mini-APEX-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 MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

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 MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use MrFuzzihead/Nex-N2.5-mini-APEX-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16

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 "MrFuzzihead/Nex-N2.5-mini-APEX-GGUF:BF16" \
  --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"

Ghost MTP head

by IsValorum - opened 9 days ago

Discussion

IsValorum

9 days ago

Hi @MrFuzzihead ,

Thanks for providing these APEX quants!

I ran into a model loading error in llama.cpp / llama-server (and LM Studio):
error loading model: check_tensor_dims: tensor 'blk.40.attn_norm.weight' not found

Root Cause

The base nex-agi/Nex-N2.5-mini model card includes mtp_num_hidden_layers: 1 in config.json, but no actual MTP tensors were exported in the base weights.

During GGUF conversion:

qwen35moe.block_count was written as 41 (40 base layers + 1 MTP layer).
qwen35moe.nextn_predict_layers was written as 1.

However, the tensor payload only contains blocks blk.0 through blk.39 (40 blocks total). When llama.cpp tries to load the model, it expects blk.40 because block_count == 41, causing llama-server to fail.

Solution / Patch

Patching the GGUF header fields:

qwen35moe.block_count: 41 ➔ 40
qwen35moe.nextn_predict_layers: 1 ➔ 0

Fixes the issue immediately and allows llama.cpp / LM Studio to load the GGUF without errors. Could you update the GGUF metadata headers in the repository so other users don't hit this error? Thanks!

MrFuzzihead

Owner 9 days ago

Yeah the I-Quality I used a non mtp bf16 gguf to quantize it but the rest all use with mtp "added", except mtp didn't ship with Nex. It's a known issue and I'll reupload these without the mtp included

MrFuzzihead

Owner 9 days ago

Correct quant ladder files now uploaded

MrFuzzihead changed discussion status to closed 9 days ago

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment