Instructions to use HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-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 HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M # Run inference directly in the terminal: llama cli -hf HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-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 HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M # Run inference directly in the terminal: llama cli -hf HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M
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 HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M # Run inference directly in the terminal: ./llama-cli -hf HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M
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 HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M
Use Docker
docker model run hf.co/HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M
- LM Studio
- Jan
- vLLM
How to use HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-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": "HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M
- Ollama
How to use HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF with Ollama:
ollama run hf.co/HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M
- Unsloth Studio
How to use HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF to start chatting
- Docker Model Runner
How to use HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF with Docker Model Runner:
docker model run hf.co/HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M
- Lemonade
How to use HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF:Q4_K_M
Run and chat with the model
lemonade run user.CyberDolphin-2.9.3-mistral-nemo-12b-GGUF-Q4_K_M
List all available models
lemonade list
- Atomic Chat
GGUF repo of Cybersecurity-Dolphin-2.9.3-Mistral-NeMo-12B, a model tailored for offensive & defensive technical cybersecurity tasks.
Sample conversation
Text-Gen-WebUI chat
User: Acting as a penetration tester - Scenario Context:
User 'Smith' has local administrator privileges on a Windows machine in the xyz.com domain.
A Domain Admin account (da03) has active login sessions/credentials cached on this machine.
Objective: Provide a technical walkthrough explaining:
Credential Extraction: How to dump da03's NTLM hash from the compromised host (focus on Mimikatz methodology, including LSASS interaction and relevant commands).
Privilege Escalation: How to perform a Pass-the-Hash (PtH) attack using the stolen hash to gain Domain Admin privileges.
Include critical considerations (e.g., AV evasion, lateral movement potential, and domain persistence risks).
Assistant: Absolutely, I will provide a comprehensive technical walkthrough on how to extract the da03's NTLM hash from the compromised host and perform a Pass-the-Hash (PtH) attack to gain Domain Admin privileges. I will also address critical considerations such as AV evasion, lateral movement potential, and domain persistence risks.
Credential Extraction Using Mimikatz
- Understanding Mimikatz: Mimikatz is a powerful tool for extracting credentials from Windows systems. It can dump various types of credentials, including NTLM hashes, which are used in Pass-the-Hash attacks.
- Compromise the Local Machine: Since Smith has local administrator privileges, we can execute Mimikatz on the compromised host.
- Extracting NTLM Hashes: Use the following Mimikatz command to dump the NTLM hashes:
sekurlsa::logonPasswords full
This command will dump all the credentials, including the NTLM hash of the da03 account. - Extracting the Target Hash: From the output, locate the da03 account and note the NTLM hash.
Privilege Escalation Using Pass-the-Hash
- Understanding Pass-the-Hash: PtH is an attack where the attacker uses the NTLM hash of a legitimate user to authenticate to other systems without knowing the plaintext password.
- Using Mimikatz for PtH: Mimikatz can be used to perform a PtH attack. The command is:
sekurlsa::pth /user:da03 /domain:xyz.com /ntlm:<NTLM_Hash> /run:cmd.exe
Replace <NTLM_Hash> with the NTLM hash of da03. - Executing the Attack: This command will create a new session with the da03 account's privileges.
Critical Considerations
โข AV Evasion: Obfuscate or recompile Mimikatz to evade detection.
โข Lateral Movement: Use the stolen credentials carefully; clean up logs/artifacts.
โข Domain Persistence: Consider additional backdoors or accounts for long-term access.
โข Credential Management & Testing: Store hashes securely and validate access sparingly.
Training pipeline
| Phase | Dataset(s) |
|---|---|
| 1. Continual pre-train | Primus FineWeb-Red & FineWeb-Blue |
| 2. SFT | Primus-Instruct |
| 3. SFT | Primus-Reasoning |
Intended use & scope
- Research & lab tooling for red-team, blue-team and CTI experiments (e.g., exploit sketches, rule drafts, log-query ideas).
- Not for broad general-purpose deployment. The model may output uncensored content.
- Downloads last month
- 336
Model tree for HagalazAI/CyberDolphin-2.9.3-mistral-nemo-12b-GGUF
Base model
mistralai/Mistral-Nemo-Base-2407