Classify LIGO data

Repository associated with my Master's thesis, entitled "Deep Learning for the Classification of Transient Noise and Signals from the LIGO Detectors".

The main goal is to successfully separate the 22 classes of the Gravity Spy dataset. The models trained on the dataset are also evaluated with real gravitational wave events.

The main architectures which are used are Convolutional Neural Networks, especially from the ResNet and ConvNeXt families.

The raw results can be found at https://wandb.ai/ogait/thesis_gravity_spy/sweeps.

Name		Name	Last commit message	Last commit date
Latest commit History 263 Commits
configs		configs
logs		logs
.gitignore		.gitignore
00_dset_creation.ipynb		00_dset_creation.ipynb
01_dset_stats.ipynb		01_dset_stats.ipynb
01_exploration.ipynb		01_exploration.ipynb
02_mixed_precision_figs.ipynb		02_mixed_precision_figs.ipynb
03_baseline_figs.ipynb		03_baseline_figs.ipynb
03_baseline_figs_best_views.ipynb		03_baseline_figs_best_views.ipynb
03_view_figs.ipynb		03_view_figs.ipynb
04_baseline_best_train_curve_figs.ipynb		04_baseline_best_train_curve_figs.ipynb
04_baseline_best_val_evaluation.ipynb		04_baseline_best_val_evaluation.ipynb
04_baseline_correct_val_eval.ipynb		04_baseline_correct_val_eval.ipynb
05_optimized_scratch_best_table.ipynb		05_optimized_scratch_best_table.ipynb
05_optimized_scratch_val_curve_figs.ipynb		05_optimized_scratch_val_curve_figs.ipynb
06_compare_optimized_best.ipynb		06_compare_optimized_best.ipynb
07_compare_weighted.ipynb		07_compare_weighted.ipynb
07_weighted_loss_confusion.ipynb		07_weighted_loss_confusion.ipynb
07_weighted_loss_methods.ipynb		07_weighted_loss_methods.ipynb
08_compare_focal.ipynb		08_compare_focal.ipynb
09_baseline_transfer.ipynb		09_baseline_transfer.ipynb
10_transfer_models_compare.ipynb		10_transfer_models_compare.ipynb
10_transfer_models_sizes.ipynb		10_transfer_models_sizes.ipynb
11_optimized_transfer_best_table.ipynb		11_optimized_transfer_best_table.ipynb
12_compare_best_transfer_paper.ipynb		12_compare_best_transfer_paper.ipynb
12_compare_transfer_best.ipynb		12_compare_transfer_best.ipynb
12_transfer_best_val_evaluation.ipynb		12_transfer_best_val_evaluation.ipynb
13_optimized_transfer_augs_best_table.ipynb		13_optimized_transfer_augs_best_table.ipynb
14_compare_transfer_aug_best.ipynb		14_compare_transfer_aug_best.ipynb
15_baseline_final_eval.ipynb		15_baseline_final_eval.ipynb
15_tl_best5_final_eval.ipynb		15_tl_best5_final_eval.ipynb
15_tl_fast1_final_eval.ipynb		15_tl_fast1_final_eval.ipynb
16_paper_metrics_calc.ipynb		16_paper_metrics_calc.ipynb
17_gw_dset_stats.ipynb		17_gw_dset_stats.ipynb
17_real_gw_dataset.ipynb		17_real_gw_dataset.ipynb
17_real_gw_samples.ipynb		17_real_gw_samples.ipynb
18_baseline_gw_eval_O3.ipynb		18_baseline_gw_eval_O3.ipynb
18_baseline_gw_eval_v1.ipynb		18_baseline_gw_eval_v1.ipynb
18_baseline_gw_eval_v2.ipynb		18_baseline_gw_eval_v2.ipynb
18_baseline_gw_eval_v3.ipynb		18_baseline_gw_eval_v3.ipynb
18_baseline_gw_eval_v4.ipynb		18_baseline_gw_eval_v4.ipynb
18_baseline_gw_eval_v5.ipynb		18_baseline_gw_eval_v5.ipynb
18_baseline_gw_eval_v6.ipynb		18_baseline_gw_eval_v6.ipynb
18_tl_best5_gw_eval.ipynb		18_tl_best5_gw_eval.ipynb
18_tl_best5_gw_eval_O3.ipynb		18_tl_best5_gw_eval_O3.ipynb
18_tl_fast1_gw_eval.ipynb		18_tl_fast1_gw_eval.ipynb
19_model_archs.ipynb		19_model_archs.ipynb
20_O3_glitches_exploration.ipynb		20_O3_glitches_exploration.ipynb
21_baseline_glitchesO3_eval.ipynb		21_baseline_glitchesO3_eval.ipynb
21_baseline_glitchesO3_eval_v2.ipynb		21_baseline_glitchesO3_eval_v2.ipynb
21_o3gws_paper_figs.ipynb		21_o3gws_paper_figs.ipynb
21_tl_best5_glitchesO3_eval.ipynb		21_tl_best5_glitchesO3_eval.ipynb
21_tl_best5_glitchesO3_eval_v2.ipynb		21_tl_best5_glitchesO3_eval_v2.ipynb
22_imgs_samuel.ipynb		22_imgs_samuel.ipynb
23_O3_glitches_labels_analysis.ipynb		23_O3_glitches_labels_analysis.ipynb
51_compare_normalized.ipynb		51_compare_normalized.ipynb
61_model_soups.ipynb		61_model_soups.ipynb
91_grad_cams.ipynb		91_grad_cams.ipynb
LICENSE		LICENSE
README.md		README.md
generate_gspy_O3.py		generate_gspy_O3.py
generate_real_gw_samples.py		generate_real_gw_samples.py
generate_real_gw_samples_v2.py		generate_real_gw_samples_v2.py
generate_real_gw_samples_v3.py		generate_real_gw_samples_v3.py
gspy_dset.py		gspy_dset.py
gw_dset.py		gw_dset.py
model_eval.py		model_eval.py
my_utils.py		my_utils.py
run_names.sh		run_names.sh
run_numbers.sh		run_numbers.sh
wandb_sweep.py		wandb_sweep.py

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Repository files navigation

Classify LIGO data

About

Uh oh!

Languages

License

tttiago/GravitySpyClassifier

Folders and files

Latest commit

History

Repository files navigation

Classify LIGO data

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Languages