Modelpublic

jackal79/tle-orbit-explainer

sourceHugging Faceotherupdated 1y agoView on Hugging Face

0likes7downloads

Model Card

tle-orbit-explainer

A LoRA adapter for Qwen-1.5-7B that translates raw Two-Line Elements (TLEs) into natural-language orbit explanations, decay risk scores, and anomaly flags for general space awareness workflows.

Model Details

Model Description


Developed by	Jack Al-Kahwati / Stardrive
Funded by	⬜️ (Self-funded)
Shared by	jackal79 (Hugging Face)
Model type	LoRA adapter (`peft==0.10.0`)
Languages	English
License	TLE-Orbit-NonCommercial v1.0 (custom terms)
Finetuned from	`Qwen/Qwen1.5-7B`

Model Sources


Repository	https://huggingface.co/jackal79/tle-orbit-explainer
Paper / Blog	https://medium.com/@jack_16944/enhancing-space-awareness-with-fine-tuned-transformer-models-introducing-tle-orbit-explainer-67ae40653ed5

Uses

Direct Use

—Quick summarization of satellite orbital states for analysts
—Plain-language TLE explanations for educational purposes
—Offline dataset labeling (orbital classifications)

Downstream Use

—Combine with SGP4 for enhanced position forecasting
—Integration into satellite autonomy stacks (cubesats, small-scale hardware)
—Pre-prompted agent support in secure orbital management workflows

Out-of-Scope Use

—High-precision orbit propagation without additional physics modeling
—Applications related to targeting, weapons systems, or lethal autonomous decisions
—Jurisdictions prohibiting ML or data export (verify with ITAR/EAR guidelines)

Bias, Risks, & Limitations

Category	Note
Data bias	Trained primarily on decayed objects (`DECAY = 1`), possibly underestimating longevity for active satellites.
Temporal limits	Operates on snapshot data; does not handle continuous high-frequency time-series.
Language	Supports explanations in English only.
Accuracy	Potential inaccuracies in decay date predictions; verify independently.

Recommendations

Incorporate independent physics-based validation before operational use and maintain a human-in-the-loop for any critical or high-risk decisions.

How to Get Started

python

from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
from peft import PeftModel

base = "Qwen/Qwen1.5-7B"
lora = "jackal79/tle-orbit-explainer"

tok   = AutoTokenizer.from_pretrained(base)
model = AutoModelForCausalLM.from_pretrained(base, device_map="auto")
model = PeftModel.from_pretrained(model, lora)  # merges LoRA

pipe = pipeline("text-generation", model=model, tokenizer=tok, device=0)

prompt = """### Prompt:
1 25544U 98067A   24079.07757601 .00016717 00000+0 10270-3 0  9994
2 25544  51.6400 337.6640 0007776  35.5310 330.5120 15.50377579499263

### Reasoning:
"""
print(pipe(prompt, max_new_tokens=120)[0]["generated_text"])

License

This model is released under the TLE-Orbit-NonCommercial License v1.0.

—✅ Free for non-commercial use, research, and internal evaluation
—🚫 Commercial, operational, or for-profit use requires a separate license

To request a commercial license, contact: jack@thestardrive.com