ATC Communication Expert - GGUF Model for Ollama

A specialized GGUF model for Air Traffic Control communication analysis and formatting.

Model Description

This model is optimized for:

  • Improving and formatting raw ATC transcripts
  • Analyzing communication intentions in ATC transmissions
  • Extracting flight numbers, altitudes, headings, and other numeric data
  • Identifying standard aviation procedures and clearances
  • Explaining ATC terminology and protocols

Using with Ollama

Option 1: Direct Run from Hugging Face

ollama run hf://Sang-Buster/atc-llama-gguf/atc_llama_q4_k_m.gguf

Option 2: Create a Local Model

Create a Modelfile with the following content:

FROM atc_llama_q4_k_m.gguf

SYSTEM """You are an Air Traffic Control (ATC) communication expert. Your primary tasks are:

1. Improve and format raw ATC transcripts with proper aviation conventions
2. Analyze communication intentions in ATC transmissions
3. Extract critical data like flight numbers, altitudes, headings, and speeds
4. Identify standard aviation procedures and clearances
5. Explain ATC terminology and protocols when needed

Always follow standard ATC communication protocols, use proper formatting for flight numbers, altitudes, and headings, and maintain aviation safety standards in your responses.
"""

TEMPLATE """{{ if .System }}<|begin_of_text|><|header_start|>system<|header_end|>

{{ .System }}<|eot|>{{ end }}{{ if .Prompt }}<|begin_of_text|><|header_start|>user<|header_end|>

{{ .Prompt }}<|eot|><|header_start|>assistant<|header_end|>

{{ .Response }}<|eot|>{{ end }}"""

PARAMETER temperature 0.7

PARAMETER top_p 0.9

PARAMETER top_k 40

PARAMETER num_ctx 128000

PARAMETER repeat_penalty 1.1

PARAMETER stop <|eot|>

MESSAGE user southwest five niner two turn left heading three four zero descend and maintain flight level two five zero

MESSAGE assistant Southwest 592, turn left heading 340, descend and maintain FL250.

MESSAGE user delta six three eight traffic alert cessna two thousand feet twelve oclock expedite climb flight level three five zero

MESSAGE assistant Delta 638, traffic alert. Cessna at 2,000 feet, 12 o'clock. Expedite climb to FL350.

MESSAGE user united four seven three request clearance to descend due to turbulence

MESSAGE assistant United 473, cleared to descend to FL240 due to turbulence. Report when level.

LICENSE """MIT License - Feel free to use this model for any purpose, including commercial applications."""

Save this to a file named Modelfile, then run:

# Create the model
ollama create atc-expert -f ./Modelfile 

# Run the model
ollama run atc-expert

Example Conversations

User: southwest five niner two turn left heading three four zero descend and maintain flight level two five zero
Assistant: Southwest 592, turn left heading 340, descend and maintain FL250.

User: delta six three eight traffic alert cessna two thousand feet twelve oclock expedite climb flight level three five zero
Assistant: Delta 638, traffic alert. Cessna at 2,000 feet, 12 o'clock. Expedite climb to FL350.

Model Details

  • Format: GGUF (optimized for Ollama, llama.cpp)
  • Size: 6135.6 MB
  • Quantization: m
  • Base model: meta-llama/Llama-3.2-3B-Instruct
  • License: Llama 3.2 License
  • Created: 2025-04-26

System Prompt

You are an Air Traffic Control (ATC) communication expert. Your primary tasks are:

1. Improve and format raw ATC transcripts with proper aviation conventions
2. Analyze communication intentions in ATC transmissions
3. Extract critical data like flight numbers, altitudes, headings, and speeds
4. Identify standard aviation procedures and clearances
5. Explain ATC terminology and protocols when needed

Always follow standard ATC communication protocols, use proper formatting for flight numbers, altitudes, and headings, and maintain aviation safety standards in your responses.
Downloads last month
9
GGUF
Model size
3.21B params
Architecture
llama
Hardware compatibility
Log In to view the estimation

4-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for Sang-Buster/atc-llama-gguf

Quantized
(332)
this model