Instructions to use my-ai-stack/Stack-3.0-Omni-Nexus with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use my-ai-stack/Stack-3.0-Omni-Nexus with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="my-ai-stack/Stack-3.0-Omni-Nexus")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("my-ai-stack/Stack-3.0-Omni-Nexus")
model = AutoModelForCausalLM.from_pretrained("my-ai-stack/Stack-3.0-Omni-Nexus", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use my-ai-stack/Stack-3.0-Omni-Nexus 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 my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0
# Run inference directly in the terminal:
llama cli -hf my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0
# Run inference directly in the terminal:
llama cli -hf my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0

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 my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0
# Run inference directly in the terminal:
./build/bin/llama cli -hf my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0

Use Docker

docker model run hf.co/my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0

Llama new
LM Studio
Jan

vLLM

How to use my-ai-stack/Stack-3.0-Omni-Nexus with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "my-ai-stack/Stack-3.0-Omni-Nexus"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "my-ai-stack/Stack-3.0-Omni-Nexus",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0

SGLang

How to use my-ai-stack/Stack-3.0-Omni-Nexus with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "my-ai-stack/Stack-3.0-Omni-Nexus" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "my-ai-stack/Stack-3.0-Omni-Nexus",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "my-ai-stack/Stack-3.0-Omni-Nexus" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "my-ai-stack/Stack-3.0-Omni-Nexus",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use my-ai-stack/Stack-3.0-Omni-Nexus with Ollama:
```
ollama run hf.co/my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0
```
Unsloth Desktop

How to use my-ai-stack/Stack-3.0-Omni-Nexus with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0

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": "my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use my-ai-stack/Stack-3.0-Omni-Nexus with Docker Model Runner:
```
docker model run hf.co/my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0
```

Lemonade

How to use my-ai-stack/Stack-3.0-Omni-Nexus with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0

Run and chat with the model

lemonade run user.Stack-3.0-Omni-Nexus-Q8_0

List all available models

lemonade list

Hermes Agent

How to use my-ai-stack/Stack-3.0-Omni-Nexus with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0

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 my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use my-ai-stack/Stack-3.0-Omni-Nexus with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0

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 "my-ai-stack/Stack-3.0-Omni-Nexus:Q8_0" \
  --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"

Walid Sobhi commited on Apr 24

Commit

dfd8515

verified ·

1 Parent(s): e9a0edb

Update README: add benchmark results, GGUF file

Browse files

Files changed (1) hide show

README.md +79 -64

README.md CHANGED Viewed

@@ -3,81 +3,96 @@ license: apache-2.0
 language:
 - en
 - ar
-library_name: gguf
 tags:
 - enterprise
 - sovereign-ai
 - global-intelligence
 - high-parameter
 - omni-nexus
 ---
-<div style="background-color: #05070a; padding: 60px 30px; border-radius: 32px; border: 1px solid #1e293b; font-family: -apple-system, system-ui, sans-serif; color: #ffffff; text-align: center; margin-top: 20px; margin-bottom: 40px; box-shadow: 0 40px 80px rgba(0,0,0,0.9);">
-  <div align="center" style="margin-bottom: 40px;">
-    <div style="width: 120px; height: 12px; background: #db2777; border-radius: 20px; box-shadow: 0 0 30px rgba(219, 39, 119, 0.7); margin-bottom: 10px;"></div>
-    <div style="width: 120px; height: 12px; background: #db2777; border-radius: 20px; margin-bottom: 10px; opacity: 0.6;"></div>
-    <div style="width: 120px; height: 12px; background: #c084fc; border-radius: 20px; margin-bottom: 10px; opacity: 0.4;"></div>
-    <div style="width: 120px; height: 12px; background: #ffffff; border-radius: 20px; opacity: 0.1;"></div>
-  </div>
-  <h1 style="font-size: 4.5em; font-weight: 900; margin: 0; color: #ffffff; letter-spacing: -4px; line-height: 0.85;">
-    STACK AI
-  </h1>
-  <p style="font-size: 1.2em; color: #db2777; font-weight: 800; text-transform: uppercase; letter-spacing: 8px; margin-top: 25px;">
-    OMNI NEXUS <span style="color: #ffffff;">HEAVY</span> / 14GB
-  </p>
-  <div style="background: linear-gradient(90deg, transparent, #db2777, transparent); height: 1px; width: 60%; margin: 40px auto;"></div>
-  <p style="font-size: 1.5em; color: #cbd5e1; max-width: 750px; margin: 0 auto; line-height: 1.4; font-weight: 200;">
-    The <b style="color: #ffffff; font-weight: 600;">Full-Spectrum Intelligence Layer</b>. Expanded parameter density for uncompromised global reasoning and complex strategic synthesis.
-  </p>
-  <div style="margin-top: 45px; display: flex; flex-wrap: wrap; justify-content: center; gap: 15px;">
-    <div style="background: #111827; border: 1px solid #db2777; padding: 12px 24px; border-radius: 12px; color: #db2777; font-size: 11px; font-weight: 900; letter-spacing: 1px;">
-      14GB DENSE WEIGHTS
-    </div>
-    <div style="background: #111827; border: 1px solid #c084fc; padding: 12px 24px; border-radius: 12px; color: #c084fc; font-size: 11px; font-weight: 900; letter-spacing: 1px;">
-      HIGH-FIDELITY GGUF
-    </div>
-    <div style="background: #111827; border: 1px solid #ffffff; padding: 12px 24px; border-radius: 12px; color: #ffffff; font-size: 11px; font-weight: 900; letter-spacing: 1px;">
-      GLOBAL MULTILINGUAL
-    </div>
-  </div>
-  <div style="margin-top: 60px; display: flex; flex-wrap: wrap; justify-content: center; gap: 25px; text-align: left;">
-    <div style="flex: 1; min-width: 300px; max-width: 420px; background: rgba(219, 39, 119, 0.03); padding: 35px; border-radius: 24px; border: 1px solid rgba(219, 39, 119, 0.2);">
-      <h3 style="color: #ffffff; margin-top: 0; font-size: 1.4em;">🧠 Advanced Reasoning</h3>
-      <p style="font-size: 15px; color: #94a3b8; line-height: 1.8;">Utilizes the expanded 14GB weight profile for superior nuance in technical documentation, legal drafting, and global market analysis.</p>
-    </div>
-    <div style="flex: 1; min-width: 300px; max-width: 420px; background: rgba(192, 132, 252, 0.03); padding: 35px; border-radius: 24px; border: 1px solid rgba(192, 132, 252, 0.2);">
-      <h3 style="color: #ffffff; margin-top: 0; font-size: 1.4em;">🛡️ Sovereign Hardening</h3>
-      <p style="font-size: 15px; color: #94a3b8; line-height: 1.8;">Maximum parameter retention for air-gapped deployments. Optimized for enterprise clusters and high-VRAM workstation environments.</p>
-    </div>
-  </div>
-  <div style="margin-top: 60px; background: #000000; padding: 30px; border-radius: 20px; border: 1px solid #1e293b; display: inline-block; width: 100%; max-width: 600px; text-align: left;">
-    <p style="color: #475569; font-size: 12px; margin-bottom: 15px; font-family: monospace;">// Deploy High-Density Sovereign Layer</p>
-    <code style="color: #db2777; font-size: 16px;">$ ollama run stack-3.0-nexus-heavy</code>
-  </div>
-  <p style="margin-top: 60px; color: #334155; font-size: 10px; letter-spacing: 4px; text-transform: uppercase; font-weight: 700;">
-    © 2026 MY-AI-STACK GLOBAL INFRASTRUCTURE
-  </p>
 </div>
-## 🛠 Model Technicalities (Heavy Edition)
-Omni Nexus Heavy (14GB) is designed for environments where **precision outweighs portability**.
-### Requirements
-- **VRAM Recommendation:** 16GB+ for optimal inference speed.
-- **Architecture:** GGUF (Optimized for Llamacpp / Ollama).
-- **Use Case:** Complex agentic workflows, long-context technical reasoning, and high-fidelity multilingual translation.
 ```bash
-# Agentic Configuration
-export ANTHROPIC_BASE_URL="http://localhost:11434/v1"
-claude --model stack-3.0-nexus-heavy

 language:
 - en
 - ar
+- code
+library_name: transformers
 tags:
 - enterprise
 - sovereign-ai
 - global-intelligence
 - high-parameter
 - omni-nexus
+- qwen
+- code-generation
+- reasoning
+- gguf
+- llama-cpp
+base_model: Qwen/Qwen2.5-Coder-7B
 ---
+# STACK AI — Omni Nexus Alpha (Q8 GGUF)
+<div style="background-color: #05070a; padding: 40px 20px; border-radius: 20px; border: 1px solid #1e293b; text-align: center; margin: 20px 0;">
+**OMNI NEXUS ALPHA** · Q8 GGUF (7.6GB)
+Full-Spectrum Intelligence Layer. Built on Qwen2.5-Coder-7B with Stack-3.0 training for enhanced reasoning, code generation, and multilingual capabilities.
 </div>
+## Benchmark Results
+Evaluated on Google Cloud Tesla V100 16GB via Ollama inference engine.
+| Benchmark | Score | Std Dev |
+|-----------|-------|---------|
+| MMLU (5-shot) | **59.89%** | — |
+| ARC-Challenge (0-shot) | **83.28%** | ±1.09% |
+| HellaSwag (0-shot) | **59.61%** | — |
+| GSM8K (8-shot) | **52.39%** | — |
+| Winogrande (0-shot) | **52.01%** | — |
+| TruthfulQA (0-shot) | **45.04%** | — |
+| HumanEval (0-shot) | **85.37%** | — |
+| MBPP (3-shot) | **79.80%** | — |
+## Model Details
+| Property | Value |
+|----------|-------|
+| Base Model | Qwen/Qwen2.5-Coder-7B-Instruct |
+| Training | Stack-3.0 LoRA fine-tune (1,875 steps) |
+| Quantization | Q8_0 GGUF |
+| File Size | 7.6GB |
+| Framework | Ollama / llama.cpp |
+## Quick Start — Ollama
 ```bash
+# Install Ollama, then:
+ollama run my-ai-stack.com/omni-nexus-alpha-q8
+# Or pull the GGUF directly
+# (see GGUF section below)
+```
+## GGUF File
+The quantized model file: `Omni-Nexus-Alpha-Q8_0.gguf`
+Use with llama.cpp:
+```bash
+# Quantize (if you have the FP16)
+./llama-quantize ./Omni-Nexus-Alpha-Q8_0.gguf Q8_0
+# Run inference
+./llama-cli -m Omni-Nexus-Alpha-Q8_0.gguf -p "Your prompt here"
+```
+## Model Card
+- **Base Model**: [Qwen/Qwen2.5-Coder-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-Coder-7B-Instruct)
+- **Training Framework**: Stack-3.0 (custom training pipeline)
+- **Evaluation**: lm-evaluation-harness + custom chat-API benchmarks
+- **Hardware**: Google Cloud Tesla V100 16GB
+## Notes
+- This is the **Q8 GGUF quantized** version (7.6GB). A full FP16 model (~14GB) is available in the same repository.
+- Benchmark scores obtained via direct Ollama chat completions API (Ollama does not expose per-token logprobs).
+- Winogrande and HellaSwag via lm_eval harness excluded due to proxy logprobs being non-informative (~random).
+---
+<p style="text-align: center; color: #64748b; margin-top: 40px;">
+Built by <a href="https://huggingface.co/my-ai-stack">My AI Stack</a> · © 2026
+</p>