SAR-predict-Hs

Code of internship: Deep learning inversion of sea surface processes from SAR imaging

Content of the repository:

Preprocessing data: Convert h5 files to .pt files to load data into RAM to avoid frequent hard disk reading (I/O will limit performance)
Preliminary experiments: Reproduce of Quach et al.(2020)
Attention-based network: Training of proposed attention-based networks
Attention-based network - real part: Training of proposed attention-based network without imaginary part
Attention-based network - data augmentation: Training of proposed attention-based network with data augmentation
Increase channels and CNNs: Training of increasing channel of Quach et al.(2020) and the model proposed in this study without any high level features

Name		Name	Last commit message	Last commit date
Latest commit History 21 Commits
Attention-based network - data augmentation		Attention-based network - data augmentation
Attention-based network - real part		Attention-based network - real part
Attention-based network		Attention-based network
Image		Image
Importance analysis		Importance analysis
Only increase channel and CNN		Only increase channel and CNN
Preliminary experiments		Preliminary experiments
Preprocessing data		Preprocessing data
LICENSE		LICENSE
README.md		README.md