Renugadevi82/cisco-nx-ai-gguf-f16

Cisco NX-AI GGUF F16

16-bit GGUF (highest quality, largest size)

Model Information

• Base Model: TinyLlama-1.1B
• Fine-tuned for: Cisco network configuration
• Quantization: F16
• File size: 2.05 GB

Download

# Using wget
wget https://huggingface.co/Renugadevi82/cisco-nx-ai-gguf-f16/resolve/main/cisco-nx-ai-f16.gguf

# Using curl
curl -L https://huggingface.co/Renugadevi82/cisco-nx-ai-gguf-f16/resolve/main/cisco-nx-ai-f16.gguf -o cisco-nx-ai-f16.gguf

# Using Hugging Face CLI
huggingface-cli download Renugadevi82/cisco-nx-ai-gguf-f16 cisco-nx-ai-f16.gguf --local-dir .

Usage with llama.cpp

# Interactive mode
./main -m cisco-nx-ai-f16.gguf -i -n 256 --color -r "User:" -f prompts/chat-with-cisco.txt

# Single prompt
./main -m cisco-nx-ai-f16.gguf -p "Configure VLAN 100 with name Management" -n 100

# With specific parameters
./main -m cisco-nx-ai-f16.gguf \
  --temp 0.7 \
  --top-k 40 \
  --top-p 0.9 \
  --repeat-penalty 1.1 \
  -p "Show running configuration for interface GigabitEthernet0/1"

Usage with Python (llama-cpp-python)

from llama_cpp import Llama

# Load model
llm = Llama(
    model_path="cisco-nx-ai-f16.gguf",
    n_ctx=512,
    n_threads=4,
    n_gpu_layers=35  # Adjust based on your GPU memory
)

# Generate response
prompt = "Configure OSPF on area 0 with router ID 1.1.1.1"
response = llm(
    prompt,
    max_tokens=100,
    temperature=0.7,
    top_p=0.9,
    echo=True
)

print(response['choices'][0]['text'])

Usage with LangChain

from langchain.llms import LlamaCpp
from langchain.prompts import PromptTemplate
from langchain.chains import LLMChain

# Initialize model
llm = LlamaCpp(
    model_path="cisco-nx-ai-f16.gguf",
    temperature=0.7,
    max_tokens=100,
    n_ctx=512
)

# Create prompt template
template = """You are a Cisco network configuration assistant.

Task: {task}

Response:"""

prompt = PromptTemplate(template=template, input_variables=["task"])
chain = LLMChain(prompt=prompt, llm=llm)

# Generate
result = chain.run("Configure a static route to 10.0.0.0/24 via 192.168.1.1")
print(result)

Model Capabilities

This model is fine-tuned for Cisco networking tasks including:

• VLAN configuration
• Interface configuration
• Routing protocols (OSPF, BGP, EIGRP)
• Access control lists (ACLs)
• Network troubleshooting commands

Recommended Settings

• Temperature: 0.7-0.8
• Top-K: 40
• Top-P: 0.9
• Repeat Penalty: 1.1
• Context Length: 512

License

Apache 2.0