File size: 1,497 Bytes
6d90380
 
 
 
 
 
 
 
 
 
3c37bad
6d90380
 
 
 
 
 
 
 
3c37bad
aa52f48
3c37bad
aa52f48
 
6d90380
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
---
license: apache-2.0
tags:
  - flight-planning
  - transformer
  - coordinate-prediction
  - sequence-to-sequence
  - count-classification
---
# Flight Plan Coordinate Prediction Model (Seq2SeqCoordsTransformer)
The encoder-decoder transformer model was trained for an AI flight planning project. Predicts normalized coordinates directly and waypoint count via classification.
## Model Description
Seq2SeqCoordsTransformer architecture using `torch.nn.Transformer`. Predicts normalized lat/lon coordinates autoregressively and waypoint count (0-10) via classification head on encoder output.
* Embed Dim: 256, Heads: 8, Enc Layers: 4, Dec Layers: 4, Max Waypoints: 10
## Intended Use
Research prototype. **Not for real-world navigation.**
## Limitations
Accuracy depends on data/tuning. Fixed max waypoints (10). Not certified. **Architecture differs significantly from previous versions in this repo.**
## How to Use
This requires loading the custom `Seq2SeqCoordsTransformer` class and weights. Generation requires autoregressive decoding and taking the argmax of the count logits.

Read this article - https://medium.com/ai-simplified-in-plain-english/building-a-transformer-model-with-seq2seq-architecture-for-flight-planning-0bdd1fecaefe


## Training Data
Trained on `frankmorales2020/flight_plan_waypoints` - https://huggingface.co/datasets/frankmorales2020/flight_plan_waypoints.
## Contact
Frank Morales, BEng, MEng, SMIEEE (Boeing ATF) - https://www.linkedin.com/in/frank-morales1964/