JasonXF commited on
Commit
db1dd03
·
verified ·
1 Parent(s): dec99c4

sync-o2-partial-artifacts

Browse files
Files changed (35) hide show
  1. .gitattributes +1 -0
  2. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/compiled/o2_data_efficiency_metrics.csv +4 -0
  3. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/compiled/o2_data_efficiency_metrics.json +116 -0
  4. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/logs/o2_gpu0_bilstm.log +13 -0
  5. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/logs/o2_gpu1_fno_transformer.log +39 -0
  6. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/logs/o2_postprocess_after_training.log +15 -0
  7. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/O2_data_efficiency_report.md +33 -0
  8. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/O2_data_efficiency_report.pdf +3 -0
  9. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/O2_data_efficiency_report.tex +42 -0
  10. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig1_mse_degradation.pdf +0 -0
  11. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig1_mse_degradation.png +3 -0
  12. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig2_r2_retention.pdf +0 -0
  13. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig2_r2_retention.png +3 -0
  14. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig3_retention_heatmap.pdf +0 -0
  15. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig3_retention_heatmap.png +3 -0
  16. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig4_training_curves.pdf +0 -0
  17. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig4_training_curves.png +3 -0
  18. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig5_results_table.pdf +0 -0
  19. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig5_results_table.png +3 -0
  20. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/o2_report_conclusions.json +5 -0
  21. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/o2_report_summary.csv +4 -0
  22. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/run_scripts/run_gpu0_bilstm_o2.sh +17 -0
  23. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/run_scripts/run_gpu1_fno_transformer_o2.sh +17 -0
  24. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/run_scripts/run_o2_postprocess_after_training.sh +26 -0
  25. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/BiLSTM/E-AC80/checkpoints/BiLSTM_E-AC80_best.pt +3 -0
  26. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/BiLSTM/E-AC80/config.json +81 -0
  27. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/BiLSTM/E-AC80/dataset_indices.pkl +3 -0
  28. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/BiLSTM/E-AC80/split_info.txt +8 -0
  29. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/BiLSTM/E-AC80/training_log.csv +12 -0
  30. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/FNO-Large/E-AC80/checkpoints/FNO-Large_E-AC80_best.pt +3 -0
  31. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/FNO-Large/E-AC80/config.json +87 -0
  32. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/FNO-Large/E-AC80/dataset_indices.pkl +3 -0
  33. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/FNO-Large/E-AC80/split_info.txt +8 -0
  34. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/FNO-Large/E-AC80/training_log.csv +38 -0
  35. experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/manifest.json +181 -0
.gitattributes CHANGED
@@ -59,3 +59,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
61
  experiments/O1_baseline/2026-06-09_xgb2_full_bs64/artifact/report/O1_baseline_report.pdf filter=lfs diff=lfs merge=lfs -text
 
 
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
61
  experiments/O1_baseline/2026-06-09_xgb2_full_bs64/artifact/report/O1_baseline_report.pdf filter=lfs diff=lfs merge=lfs -text
62
+ experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/O2_data_efficiency_report.pdf filter=lfs diff=lfs merge=lfs -text
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/compiled/o2_data_efficiency_metrics.csv ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ model,run_id,axis,train_gms_used,use_scales_count,pool_samples,train_samples,val_samples,test_samples,best_epoch,train_time_s,peak_gpu_gb,epochs,batch_size,lr,checkpoint,numel,numel_m,real_params,real_params_m,trainable,mse,rmse,mae,r2,pfa_mse_g2,pfa_rmse_g,pfa_mae_g,pfa_r2,loss,spectral_log_mse,high_freq_log_mse,latency_ms_per_sample,r2_retention,mse_ratio,time_ratio
2
+ BiLSTM,E-Base,full,3000,57,171000,136800,34200,27018,,27026.021858215332,15.61366891860962,50,64,0.0003,/data/home/jason/Seismic-FNO/6-code/O1-baseline-comparison/results/runs/BiLSTM/checkpoints/BiLSTM_best.pt,8410113,8.410113,8410113,8.410113,8410113,0.0099188614842287,0.0995934811331983,0.0432213771548134,0.9872496731015952,0.0003300793328829,0.0181680855591055,0.0108249000931961,0.987565034951701,0.0099240713056515,0.1852839740596482,0.2222294335740114,1.0494697839021685,1.0,1.0,1.0
3
+ FNO-Large,E-Base,full,3000,57,171000,136800,34200,27018,,7784.3661642074585,3.2975897789001465,50,64,0.001,6-code/O1-baseline-comparison/results/runs/FNO-Large-bs64/checkpoints/FNO-Large-bs64_best.pt,8505793,8.505793,16927169,16.927169,8505793,0.0448808919490714,0.2118511079722536,0.1050106563313501,0.9423072854931448,0.0010877546760695,0.0329811260582404,0.0204876376139049,0.9590213926454952,0.0449616025107302,0.3416421673399337,0.3977926703209573,0.1078980043530464,1.0,1.0,1.0
4
+ Transformer,E-Base,full,3000,57,171000,136800,34200,27018,,8430.805659532547,2.5364460945129395,50,64,0.0003,/data/home/jason/Seismic-FNO/6-code/O1-baseline-comparison/results/runs/Transformer/checkpoints/Transformer_best.pt,7957775,7.957775,7957775,7.957775,7957775,0.0213903603177851,0.1462544369165773,0.0660383013369173,0.9725034887360728,0.000533425337507,0.023096002630479,0.0149489011169053,0.9799044509395886,0.0214709252808199,0.3250373078083033,0.3907267726411774,0.1057809218764305,1.0,1.0,1.0
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/compiled/o2_data_efficiency_metrics.json ADDED
@@ -0,0 +1,116 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "model":"BiLSTM",
4
+ "run_id":"E-Base",
5
+ "axis":"full",
6
+ "train_gms_used":3000,
7
+ "use_scales_count":57,
8
+ "pool_samples":171000,
9
+ "train_samples":136800,
10
+ "val_samples":34200,
11
+ "test_samples":27018,
12
+ "best_epoch":"",
13
+ "train_time_s":27026.0218582153,
14
+ "peak_gpu_gb":15.6136689186,
15
+ "epochs":50,
16
+ "batch_size":64,
17
+ "lr":0.0003,
18
+ "checkpoint":"\/data\/home\/jason\/Seismic-FNO\/6-code\/O1-baseline-comparison\/results\/runs\/BiLSTM\/checkpoints\/BiLSTM_best.pt",
19
+ "numel":8410113,
20
+ "numel_m":8.410113,
21
+ "real_params":8410113,
22
+ "real_params_m":8.410113,
23
+ "trainable":8410113,
24
+ "mse":0.0099188615,
25
+ "rmse":0.0995934811,
26
+ "mae":0.0432213772,
27
+ "r2":0.9872496731,
28
+ "pfa_mse_g2":0.0003300793,
29
+ "pfa_rmse_g":0.0181680856,
30
+ "pfa_mae_g":0.0108249001,
31
+ "pfa_r2":0.987565035,
32
+ "loss":0.0099240713,
33
+ "spectral_log_mse":0.1852839741,
34
+ "high_freq_log_mse":0.2222294336,
35
+ "latency_ms_per_sample":1.0494697839,
36
+ "r2_retention":1.0,
37
+ "mse_ratio":1.0,
38
+ "time_ratio":1.0
39
+ },
40
+ {
41
+ "model":"FNO-Large",
42
+ "run_id":"E-Base",
43
+ "axis":"full",
44
+ "train_gms_used":3000,
45
+ "use_scales_count":57,
46
+ "pool_samples":171000,
47
+ "train_samples":136800,
48
+ "val_samples":34200,
49
+ "test_samples":27018,
50
+ "best_epoch":"",
51
+ "train_time_s":7784.3661642075,
52
+ "peak_gpu_gb":3.2975897789,
53
+ "epochs":50,
54
+ "batch_size":64,
55
+ "lr":0.001,
56
+ "checkpoint":"6-code\/O1-baseline-comparison\/results\/runs\/FNO-Large-bs64\/checkpoints\/FNO-Large-bs64_best.pt",
57
+ "numel":8505793,
58
+ "numel_m":8.505793,
59
+ "real_params":16927169,
60
+ "real_params_m":16.927169,
61
+ "trainable":8505793,
62
+ "mse":0.0448808919,
63
+ "rmse":0.211851108,
64
+ "mae":0.1050106563,
65
+ "r2":0.9423072855,
66
+ "pfa_mse_g2":0.0010877547,
67
+ "pfa_rmse_g":0.0329811261,
68
+ "pfa_mae_g":0.0204876376,
69
+ "pfa_r2":0.9590213926,
70
+ "loss":0.0449616025,
71
+ "spectral_log_mse":0.3416421673,
72
+ "high_freq_log_mse":0.3977926703,
73
+ "latency_ms_per_sample":0.1078980044,
74
+ "r2_retention":1.0,
75
+ "mse_ratio":1.0,
76
+ "time_ratio":1.0
77
+ },
78
+ {
79
+ "model":"Transformer",
80
+ "run_id":"E-Base",
81
+ "axis":"full",
82
+ "train_gms_used":3000,
83
+ "use_scales_count":57,
84
+ "pool_samples":171000,
85
+ "train_samples":136800,
86
+ "val_samples":34200,
87
+ "test_samples":27018,
88
+ "best_epoch":"",
89
+ "train_time_s":8430.8056595325,
90
+ "peak_gpu_gb":2.5364460945,
91
+ "epochs":50,
92
+ "batch_size":64,
93
+ "lr":0.0003,
94
+ "checkpoint":"\/data\/home\/jason\/Seismic-FNO\/6-code\/O1-baseline-comparison\/results\/runs\/Transformer\/checkpoints\/Transformer_best.pt",
95
+ "numel":7957775,
96
+ "numel_m":7.957775,
97
+ "real_params":7957775,
98
+ "real_params_m":7.957775,
99
+ "trainable":7957775,
100
+ "mse":0.0213903603,
101
+ "rmse":0.1462544369,
102
+ "mae":0.0660383013,
103
+ "r2":0.9725034887,
104
+ "pfa_mse_g2":0.0005334253,
105
+ "pfa_rmse_g":0.0230960026,
106
+ "pfa_mae_g":0.0149489011,
107
+ "pfa_r2":0.9799044509,
108
+ "loss":0.0214709253,
109
+ "spectral_log_mse":0.3250373078,
110
+ "high_freq_log_mse":0.3907267726,
111
+ "latency_ms_per_sample":0.1057809219,
112
+ "r2_retention":1.0,
113
+ "mse_ratio":1.0,
114
+ "time_ratio":1.0
115
+ }
116
+ ]
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/logs/o2_gpu0_bilstm.log ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ === O2 queue start: BiLSTM E-AC80 ===
3
+ BiLSTM E-AC80 epoch 001/50 train_mse=0.184514 val_mse=0.094526 val_r2=0.8806 peak=15.61GB
4
+ BiLSTM E-AC80 epoch 002/50 train_mse=0.088498 val_mse=0.065684 val_r2=0.9170 peak=15.61GB
5
+ BiLSTM E-AC80 epoch 003/50 train_mse=0.066572 val_mse=0.057646 val_r2=0.9272 peak=15.61GB
6
+ BiLSTM E-AC80 epoch 004/50 train_mse=0.064251 val_mse=0.057081 val_r2=0.9279 peak=15.61GB
7
+ BiLSTM E-AC80 epoch 005/50 train_mse=0.050086 val_mse=0.040376 val_r2=0.9490 peak=15.61GB
8
+ BiLSTM E-AC80 epoch 006/50 train_mse=0.037317 val_mse=0.032665 val_r2=0.9587 peak=15.61GB
9
+ BiLSTM E-AC80 epoch 007/50 train_mse=0.030899 val_mse=0.027043 val_r2=0.9658 peak=15.61GB
10
+ BiLSTM E-AC80 epoch 008/50 train_mse=0.026124 val_mse=0.023634 val_r2=0.9702 peak=15.61GB
11
+ BiLSTM E-AC80 epoch 009/50 train_mse=0.024447 val_mse=0.020375 val_r2=0.9743 peak=15.61GB
12
+ BiLSTM E-AC80 epoch 010/50 train_mse=0.021730 val_mse=0.024054 val_r2=0.9696 peak=15.61GB
13
+ BiLSTM E-AC80 epoch 011/50 train_mse=0.021081 val_mse=0.022259 val_r2=0.9719 peak=15.61GB
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/logs/o2_gpu1_fno_transformer.log ADDED
@@ -0,0 +1,39 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ === O2 queue start: FNO-Large E-AC80 ===
3
+ FNO-Large E-AC80 epoch 001/50 train_mse=0.121265 val_mse=0.085454 val_r2=0.8921 peak=3.30GB
4
+ FNO-Large E-AC80 epoch 002/50 train_mse=0.059299 val_mse=0.073123 val_r2=0.9076 peak=3.30GB
5
+ FNO-Large E-AC80 epoch 003/50 train_mse=0.046376 val_mse=0.064540 val_r2=0.9185 peak=3.30GB
6
+ FNO-Large E-AC80 epoch 004/50 train_mse=0.040007 val_mse=0.058650 val_r2=0.9259 peak=3.30GB
7
+ FNO-Large E-AC80 epoch 005/50 train_mse=0.036330 val_mse=0.054976 val_r2=0.9306 peak=3.30GB
8
+ FNO-Large E-AC80 epoch 006/50 train_mse=0.032305 val_mse=0.054311 val_r2=0.9314 peak=3.30GB
9
+ FNO-Large E-AC80 epoch 007/50 train_mse=0.030493 val_mse=0.055501 val_r2=0.9299 peak=3.30GB
10
+ FNO-Large E-AC80 epoch 008/50 train_mse=0.029097 val_mse=0.050662 val_r2=0.9360 peak=3.30GB
11
+ FNO-Large E-AC80 epoch 009/50 train_mse=0.027164 val_mse=0.050450 val_r2=0.9363 peak=3.30GB
12
+ FNO-Large E-AC80 epoch 010/50 train_mse=0.026214 val_mse=0.049896 val_r2=0.9370 peak=3.30GB
13
+ FNO-Large E-AC80 epoch 011/50 train_mse=0.025146 val_mse=0.050880 val_r2=0.9357 peak=3.30GB
14
+ FNO-Large E-AC80 epoch 012/50 train_mse=0.024458 val_mse=0.052032 val_r2=0.9343 peak=3.30GB
15
+ FNO-Large E-AC80 epoch 013/50 train_mse=0.023643 val_mse=0.048264 val_r2=0.9390 peak=3.30GB
16
+ FNO-Large E-AC80 epoch 014/50 train_mse=0.023520 val_mse=0.048628 val_r2=0.9386 peak=3.30GB
17
+ FNO-Large E-AC80 epoch 015/50 train_mse=0.022679 val_mse=0.049527 val_r2=0.9374 peak=3.30GB
18
+ FNO-Large E-AC80 epoch 016/50 train_mse=0.022033 val_mse=0.049063 val_r2=0.9380 peak=3.30GB
19
+ FNO-Large E-AC80 epoch 017/50 train_mse=0.021457 val_mse=0.047159 val_r2=0.9404 peak=3.30GB
20
+ FNO-Large E-AC80 epoch 018/50 train_mse=0.021117 val_mse=0.048310 val_r2=0.9390 peak=3.30GB
21
+ FNO-Large E-AC80 epoch 019/50 train_mse=0.020792 val_mse=0.047705 val_r2=0.9398 peak=3.30GB
22
+ FNO-Large E-AC80 epoch 020/50 train_mse=0.020154 val_mse=0.047837 val_r2=0.9396 peak=3.30GB
23
+ FNO-Large E-AC80 epoch 021/50 train_mse=0.016622 val_mse=0.046132 val_r2=0.9417 peak=3.30GB
24
+ FNO-Large E-AC80 epoch 022/50 train_mse=0.016008 val_mse=0.047560 val_r2=0.9399 peak=3.30GB
25
+ FNO-Large E-AC80 epoch 023/50 train_mse=0.015775 val_mse=0.047003 val_r2=0.9406 peak=3.30GB
26
+ FNO-Large E-AC80 epoch 024/50 train_mse=0.015629 val_mse=0.048206 val_r2=0.9391 peak=3.30GB
27
+ FNO-Large E-AC80 epoch 025/50 train_mse=0.015429 val_mse=0.046298 val_r2=0.9415 peak=3.30GB
28
+ FNO-Large E-AC80 epoch 026/50 train_mse=0.015187 val_mse=0.047017 val_r2=0.9406 peak=3.30GB
29
+ FNO-Large E-AC80 epoch 027/50 train_mse=0.015193 val_mse=0.047117 val_r2=0.9405 peak=3.30GB
30
+ FNO-Large E-AC80 epoch 028/50 train_mse=0.014857 val_mse=0.046992 val_r2=0.9407 peak=3.30GB
31
+ FNO-Large E-AC80 epoch 029/50 train_mse=0.014601 val_mse=0.046547 val_r2=0.9412 peak=3.30GB
32
+ FNO-Large E-AC80 epoch 030/50 train_mse=0.014563 val_mse=0.047840 val_r2=0.9396 peak=3.30GB
33
+ FNO-Large E-AC80 epoch 031/50 train_mse=0.014507 val_mse=0.046321 val_r2=0.9415 peak=3.30GB
34
+ FNO-Large E-AC80 epoch 032/50 train_mse=0.014403 val_mse=0.047386 val_r2=0.9402 peak=3.30GB
35
+ FNO-Large E-AC80 epoch 033/50 train_mse=0.014147 val_mse=0.046901 val_r2=0.9408 peak=3.30GB
36
+ FNO-Large E-AC80 epoch 034/50 train_mse=0.014076 val_mse=0.048018 val_r2=0.9394 peak=3.30GB
37
+ FNO-Large E-AC80 epoch 035/50 train_mse=0.014131 val_mse=0.048091 val_r2=0.9393 peak=3.30GB
38
+ FNO-Large E-AC80 epoch 036/50 train_mse=0.013998 val_mse=0.047685 val_r2=0.9398 peak=3.30GB
39
+ FNO-Large E-AC80 epoch 037/50 train_mse=0.013801 val_mse=0.046387 val_r2=0.9414 peak=3.30GB
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/logs/o2_postprocess_after_training.log ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Waiting for O2 training queues: GPU0=3914712 GPU1=3914516
2
+ 2026-06-05 10:36:13 still waiting for O2 queues...
3
+ 2026-06-05 11:06:13 still waiting for O2 queues...
4
+ 2026-06-05 11:36:13 still waiting for O2 queues...
5
+ O2 training queues finished. Collecting results...
6
+ Wrote 6-code/O2-data-efficiency/results/compiled/o2_data_efficiency_metrics.csv
7
+ Generating O2 report...
8
+ /data/home/jason/Seismic-FNO/6-code/O2-data-efficiency/scripts/generate_o2_report.py:147: UserWarning: No artists with labels found to put in legend. Note that artists whose label start with an underscore are ignored when legend() is called with no argument.
9
+ axs[-1].legend(frameon=False, fontsize=9)
10
+ /data/home/jason/Seismic-FNO/6-code/O2-data-efficiency/scripts/generate_o2_report.py:210: UserWarning: This figure includes Axes that are not compatible with tight_layout, so results might be incorrect.
11
+ fig.tight_layout()
12
+ /data/home/jason/Seismic-FNO/6-code/O2-data-efficiency/scripts/generate_o2_report.py:232: UserWarning: No artists with labels found to put in legend. Note that artists whose label start with an underscore are ignored when legend() is called with no argument.
13
+ axs[-1].legend(frameon=False, fontsize=9)
14
+ Wrote O2 report to /data/home/jason/Seismic-FNO/6-code/O2-data-efficiency/results/report
15
+ O2 postprocess complete.
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/O2_data_efficiency_report.md ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # O2 Data Efficiency Report
2
+
3
+ Generated: 2026-06-05 12:06:15
4
+
5
+ ## Protocol
6
+
7
+ The O2 study keeps the fixed 474-GM test set from O1 unchanged and prunes only the supervised train/validation pool. FNO-Large, Transformer, and BiLSTM are trained with the O1-final hyperparameters. E-Base is imported from O1 full-data results.
8
+
9
+ ## Key Conclusions
10
+
11
+ - FNO-Large: E-Base MSE=0.0449, R2=0.9423; reduced runs pending.
12
+ - Transformer: E-Base MSE=0.0214, R2=0.9725; reduced runs pending.
13
+ - BiLSTM: E-Base MSE=0.0099, R2=0.9872; reduced runs pending.
14
+
15
+ ## Figures
16
+
17
+ ![Test MSE degradation](figures/fig1_mse_degradation.png)
18
+
19
+ ![R2 retention](figures/fig2_r2_retention.png)
20
+
21
+ ![Retention heatmap](figures/fig3_retention_heatmap.png)
22
+
23
+ ![Training curves](figures/fig4_training_curves.png)
24
+
25
+ ![Results table](figures/fig5_results_table.png)
26
+
27
+ ## Quantitative Table
28
+
29
+ |Model|Run|Axis|GMs|ACs|Pool|MSE|R2|PFA RMSE|R2 Ret.|Time(h)|GPU(GB)|
30
+ |---|---|---|---|---|---|---|---|---|---|---|---|
31
+ |FNO-Large|E-Base|full|3000|57|171.0k|0.0449|0.9423|0.0330|1.0000|2.16|3.30|
32
+ |Transformer|E-Base|full|3000|57|171.0k|0.0214|0.9725|0.0231|1.0000|2.34|2.54|
33
+ |BiLSTM|E-Base|full|3000|57|171.0k|0.0099|0.9872|0.0182|1.0000|7.51|15.61|
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/O2_data_efficiency_report.pdf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da7a095d543ae6769b578905ba8503063d747dc8706c22fafab2a4d6e907558b
3
+ size 156136
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/O2_data_efficiency_report.tex ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ \documentclass[11pt]{article}
2
+ \usepackage[margin=1in]{geometry}
3
+ \usepackage{graphicx}
4
+ \usepackage{booktabs}
5
+ \usepackage{float}
6
+ \title{O2 Data Efficiency Report}
7
+ \author{Seismic-FNO Experiments}
8
+ \date{2026-06-05}
9
+ \begin{document}
10
+ \maketitle
11
+
12
+ \section*{Protocol}
13
+ The O2 study keeps the fixed 474-GM test set from O1 unchanged and prunes only
14
+ the supervised train/validation pool. FNO-Large, Transformer, and BiLSTM are
15
+ trained with the O1-final hyperparameters. E-Base is imported from O1 full-data
16
+ results.
17
+
18
+ \section*{Key Conclusions}
19
+ \begin{itemize}
20
+ \item FNO-Large: E-Base MSE=0.0449, R2=0.9423; reduced runs pending.
21
+ \item Transformer: E-Base MSE=0.0214, R2=0.9725; reduced runs pending.
22
+ \item BiLSTM: E-Base MSE=0.0099, R2=0.9872; reduced runs pending.
23
+ \end{itemize}
24
+
25
+ \section*{Figures}
26
+ \begin{figure}[H]\centering\includegraphics[width=\linewidth]{figures/fig1_mse_degradation.png}\caption{Test MSE degradation.}\end{figure}
27
+ \begin{figure}[H]\centering\includegraphics[width=\linewidth]{figures/fig2_r2_retention.png}\caption{R2 retention relative to each model's E-Base.}\end{figure}
28
+ \begin{figure}[H]\centering\includegraphics[width=\linewidth]{figures/fig3_retention_heatmap.png}\caption{Retention heatmap over GM and AC budgets.}\end{figure}
29
+ \begin{figure}[H]\centering\includegraphics[width=\linewidth]{figures/fig4_training_curves.png}\caption{Representative validation curves.}\end{figure}
30
+
31
+ \section*{Quantitative Table}
32
+ \small
33
+ \begin{tabular}{llrrrrrr}
34
+ \toprule
35
+ Model & Run & GMs & ACs & Pool & MSE & $R^2$ & $R^2$ Ret. \\
36
+ \midrule
37
+ FNO-Large & E-Base & 3000 & 57 & 171.0k & 0.0449 & 0.9423 & 1.0000 \\
38
+ Transformer & E-Base & 3000 & 57 & 171.0k & 0.0214 & 0.9725 & 1.0000 \\
39
+ BiLSTM & E-Base & 3000 & 57 & 171.0k & 0.0099 & 0.9872 & 1.0000 \\
40
+ \bottomrule
41
+ \end{tabular}
42
+ \end{document}
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig1_mse_degradation.pdf ADDED
Binary file (16.5 kB). View file
 
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig1_mse_degradation.png ADDED

Git LFS Details

  • SHA256: 5103ef2df5f0f77dc48a71fc75a48dc6f199b89e518e4f443f46e6fffe32eb45
  • Pointer size: 130 Bytes
  • Size of remote file: 57.4 kB
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig2_r2_retention.pdf ADDED
Binary file (17.7 kB). View file
 
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig2_r2_retention.png ADDED

Git LFS Details

  • SHA256: bb7487f5f98e87e2018b478bfada26e259f44bc9064b85545db796b826d2ffa7
  • Pointer size: 130 Bytes
  • Size of remote file: 68.4 kB
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig3_retention_heatmap.pdf ADDED
Binary file (22 kB). View file
 
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig3_retention_heatmap.png ADDED

Git LFS Details

  • SHA256: b5766f4e0b16ea2bbacf6029bf643c834da0d95138b9c5c9c7f7af56b13ac2e7
  • Pointer size: 130 Bytes
  • Size of remote file: 96 kB
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig4_training_curves.pdf ADDED
Binary file (15.7 kB). View file
 
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig4_training_curves.png ADDED

Git LFS Details

  • SHA256: 7e1b2bf8366b2ba308aa2a31f31e172ec2d43447563eed2728bcf8f9193d3e0f
  • Pointer size: 130 Bytes
  • Size of remote file: 63.7 kB
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig5_results_table.pdf ADDED
Binary file (20.1 kB). View file
 
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/figures/fig5_results_table.png ADDED

Git LFS Details

  • SHA256: a5636233670f7a5ba05aa6e452c617c833a1b63e40189d759670e8e22c151f49
  • Pointer size: 130 Bytes
  • Size of remote file: 60.1 kB
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/o2_report_conclusions.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ [
2
+ "FNO-Large: E-Base MSE=0.0449, R2=0.9423; reduced runs pending.",
3
+ "Transformer: E-Base MSE=0.0214, R2=0.9725; reduced runs pending.",
4
+ "BiLSTM: E-Base MSE=0.0099, R2=0.9872; reduced runs pending."
5
+ ]
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/report/o2_report_summary.csv ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ model,run_id,axis,train_gms_used,use_scales_count,pool_samples,train_samples,val_samples,test_samples,best_epoch,train_time_s,peak_gpu_gb,epochs,batch_size,lr,checkpoint,numel,numel_m,real_params,real_params_m,trainable,mse,rmse,mae,r2,pfa_mse_g2,pfa_rmse_g,pfa_mae_g,pfa_r2,loss,spectral_log_mse,high_freq_log_mse,latency_ms_per_sample,r2_retention,mse_ratio,time_ratio
2
+ FNO-Large,E-Base,full,3000,57,171000,136800,34200,27018,,7784.366164207458,3.2975897789001465,50,64,0.001,6-code/O1-baseline-comparison/results/runs/FNO-Large-bs64/checkpoints/FNO-Large-bs64_best.pt,8505793,8.505793,16927169,16.927169,8505793,0.0448808919490714,0.2118511079722536,0.1050106563313501,0.9423072854931448,0.0010877546760695,0.0329811260582404,0.0204876376139049,0.9590213926454952,0.0449616025107302,0.3416421673399337,0.3977926703209573,0.1078980043530464,1.0,1.0,1.0
3
+ Transformer,E-Base,full,3000,57,171000,136800,34200,27018,,8430.805659532547,2.5364460945129395,50,64,0.0003,/data/home/jason/Seismic-FNO/6-code/O1-baseline-comparison/results/runs/Transformer/checkpoints/Transformer_best.pt,7957775,7.957775,7957775,7.957775,7957775,0.0213903603177851,0.1462544369165773,0.0660383013369173,0.9725034887360728,0.000533425337507,0.023096002630479,0.0149489011169053,0.9799044509395886,0.0214709252808199,0.3250373078083033,0.3907267726411774,0.1057809218764305,1.0,1.0,1.0
4
+ BiLSTM,E-Base,full,3000,57,171000,136800,34200,27018,,27026.02185821533,15.61366891860962,50,64,0.0003,/data/home/jason/Seismic-FNO/6-code/O1-baseline-comparison/results/runs/BiLSTM/checkpoints/BiLSTM_best.pt,8410113,8.410113,8410113,8.410113,8410113,0.0099188614842287,0.0995934811331983,0.0432213771548134,0.9872496731015952,0.0003300793328829,0.0181680855591055,0.0108249000931961,0.987565034951701,0.0099240713056515,0.1852839740596482,0.2222294335740114,1.0494697839021685,1.0,1.0,1.0
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/run_scripts/run_gpu0_bilstm_o2.sh ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env bash
2
+ set -euo pipefail
3
+ cd /data/home/jason/Seismic-FNO
4
+ export PYTHONDONTWRITEBYTECODE=1
5
+ export CUDA_VISIBLE_DEVICES=0
6
+
7
+ /data/home/jason/miniconda3/envs/sefno/bin/python 6-code/O2-data-efficiency/scripts/run_o2_queue.py \
8
+ --models BiLSTM \
9
+ --python /data/home/jason/miniconda3/envs/sefno/bin/python \
10
+ --epochs 50 \
11
+ --batch_size 64 \
12
+ --eval_batch_size 64 \
13
+ --base_data_dir /data/home/jason/data/SesimicTransformerData \
14
+ --building_dir /data/home/jason/data/SesimicTransformerData/MDOF/knet-250/Data/fno \
15
+ --output_root 6-code/O2-data-efficiency/results \
16
+ --no_progress \
17
+ --log_interval 100
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/run_scripts/run_gpu1_fno_transformer_o2.sh ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env bash
2
+ set -euo pipefail
3
+ cd /data/home/jason/Seismic-FNO
4
+ export PYTHONDONTWRITEBYTECODE=1
5
+ export CUDA_VISIBLE_DEVICES=1
6
+
7
+ /data/home/jason/miniconda3/envs/sefno/bin/python 6-code/O2-data-efficiency/scripts/run_o2_queue.py \
8
+ --models FNO-Large Transformer \
9
+ --python /data/home/jason/miniconda3/envs/sefno/bin/python \
10
+ --epochs 50 \
11
+ --batch_size 64 \
12
+ --eval_batch_size 64 \
13
+ --base_data_dir /data/home/jason/data/SesimicTransformerData \
14
+ --building_dir /data/home/jason/data/SesimicTransformerData/MDOF/knet-250/Data/fno \
15
+ --output_root 6-code/O2-data-efficiency/results \
16
+ --no_progress \
17
+ --log_interval 100
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/run_scripts/run_o2_postprocess_after_training.sh ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/env bash
2
+ set -euo pipefail
3
+ cd /data/home/jason/Seismic-FNO
4
+ export PYTHONDONTWRITEBYTECODE=1
5
+
6
+ GPU0_PID_FILE=6-code/O2-data-efficiency/results/pids/o2_gpu0_bilstm.pid
7
+ GPU1_PID_FILE=6-code/O2-data-efficiency/results/pids/o2_gpu1_fno_transformer.pid
8
+
9
+ GPU0_PID=$(cat "$GPU0_PID_FILE")
10
+ GPU1_PID=$(cat "$GPU1_PID_FILE")
11
+
12
+ echo "Waiting for O2 training queues: GPU0=$GPU0_PID GPU1=$GPU1_PID"
13
+ while kill -0 "$GPU0_PID" 2>/dev/null || kill -0 "$GPU1_PID" 2>/dev/null; do
14
+ date '+%Y-%m-%d %H:%M:%S still waiting for O2 queues...'
15
+ sleep 1800
16
+ done
17
+
18
+ echo "O2 training queues finished. Collecting results..."
19
+ /data/home/jason/miniconda3/envs/sefno/bin/python 6-code/O2-data-efficiency/scripts/collect_o2_results.py \
20
+ --results_root 6-code/O2-data-efficiency/results
21
+
22
+ echo "Generating O2 report..."
23
+ /data/home/jason/miniconda3/envs/sefno/bin/python 6-code/O2-data-efficiency/scripts/generate_o2_report.py \
24
+ --results_root 6-code/O2-data-efficiency/results
25
+
26
+ echo "O2 postprocess complete."
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/BiLSTM/E-AC80/checkpoints/BiLSTM_E-AC80_best.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0fe8a5bcecaaf3ad346a6974ba7419e084e48d480fb36579d56bc855df0435da
3
+ size 100941869
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/BiLSTM/E-AC80/config.json ADDED
@@ -0,0 +1,81 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "BiLSTM",
3
+ "run_id": "E-AC80",
4
+ "axis": "ac",
5
+ "seed": 42,
6
+ "epochs": 50,
7
+ "batch_size": 64,
8
+ "eval_batch_size": 64,
9
+ "lr": 0.0003,
10
+ "base_lr": 0.001,
11
+ "bilstm_lr": 0.0003,
12
+ "transformer_lr": 0.0003,
13
+ "weight_decay": 0.0001,
14
+ "scheduler_step": 20,
15
+ "scheduler_gamma": 0.5,
16
+ "grad_clip": 1.0,
17
+ "train_gms_used": 3000,
18
+ "use_scales_count": 46,
19
+ "train_gms": 2400,
20
+ "val_gms": 600,
21
+ "test_gms": 474,
22
+ "train_samples": 110400,
23
+ "val_samples": 27600,
24
+ "test_samples": 27018,
25
+ "scale_indices": [
26
+ 0,
27
+ 1,
28
+ 2,
29
+ 4,
30
+ 5,
31
+ 6,
32
+ 7,
33
+ 9,
34
+ 10,
35
+ 11,
36
+ 12,
37
+ 14,
38
+ 15,
39
+ 16,
40
+ 17,
41
+ 19,
42
+ 20,
43
+ 21,
44
+ 22,
45
+ 24,
46
+ 25,
47
+ 26,
48
+ 27,
49
+ 29,
50
+ 30,
51
+ 31,
52
+ 32,
53
+ 34,
54
+ 35,
55
+ 36,
56
+ 37,
57
+ 39,
58
+ 40,
59
+ 41,
60
+ 42,
61
+ 44,
62
+ 45,
63
+ 46,
64
+ 47,
65
+ 49,
66
+ 50,
67
+ 51,
68
+ 52,
69
+ 54,
70
+ 55,
71
+ 56
72
+ ],
73
+ "split": "fixed_474_test_full_57_factors_reduced_train_val",
74
+ "params": {
75
+ "numel": 8410113,
76
+ "numel_m": 8.410113,
77
+ "real_params": 8410113,
78
+ "real_params_m": 8.410113,
79
+ "trainable": 8410113
80
+ }
81
+ }
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/BiLSTM/E-AC80/dataset_indices.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:350b63e638363a4dd6ba42f937de91a93b2ca9eac0a362d75e26696effba9c3e
3
+ size 1348786
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/BiLSTM/E-AC80/split_info.txt ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ Canonical split: fixed 474-GM test + 3000-GM pool + 80/20 train/val
2
+ Train GMs: 2400
3
+ Val GMs: 600
4
+ Test GMs: 474
5
+ Train samples: 110400
6
+ Val samples: 27600
7
+ Test samples: 27018
8
+ Scale count: 46
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/BiLSTM/E-AC80/training_log.csv ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,train_mse,train_rmse,train_mae,train_r2,val_mse,val_rmse,val_mae,val_r2,val_pfa_rmse_g,lr,train_max_grad_norm,epoch_s,gpu_peak_gb
2
+ 1,0.1845143684217665,0.4295513571411066,0.24800047843896247,0.7669326353976882,0.09452598470052083,0.3074507841923986,0.17222814416562876,0.8806167572857764,0.06258696102673016,0.0003,113.3521728515625,437.21639227867126,15.61366891860962
3
+ 2,0.08849822482344033,0.29748651200254494,0.16708426787298084,0.8882144071291098,0.0656840711628642,0.25628903831975375,0.1380961776187915,0.917043155541427,0.048389819792850536,0.0003,347.7506408691406,438.04145193099976,15.61366891860962
4
+ 3,0.06657224863098439,0.2580159852237539,0.1417162893396645,0.9159099709839427,0.05764608892321011,0.24009599939026496,0.13119099490253247,0.9271948655467708,0.04335682339086055,0.0003,205.78515625,433.01851177215576,15.61366891860962
5
+ 4,0.06425107987261049,0.2534779672330723,0.13983717831229242,0.9188419306992228,0.05708125596198483,0.23891683900885854,0.13146831765879755,0.9279082311966139,0.0415390293648752,0.0003,188.76846313476562,430.7814738750458,15.61366891860962
6
+ 5,0.05008617020906458,0.2237993972491092,0.11983992691408608,0.9367341860308024,0.04037563290213617,0.20093688785819336,0.09923721318194256,0.9490068895048542,0.03792927893725086,0.0003,24.431541442871094,450.4718918800354,15.61366891860962
7
+ 6,0.03731727703665766,0.19317680253244088,0.09860245896491451,0.9528630777996946,0.03266525762585626,0.18073532478698306,0.09113527994791666,0.9587448425760885,0.0329786106309215,0.0003,85.50286102294922,514.9481568336487,15.61366891860962
8
+ 7,0.030898608692556187,0.17578000083216574,0.08979302871188104,0.9609707505640337,0.027042742428986924,0.1644467768884113,0.08628786313264267,0.9658458963079155,0.03338080120637214,0.0003,6.544026851654053,499.80085849761963,15.61366891860962
9
+ 8,0.026124482508488898,0.16163069791499662,0.08127886158229072,0.9670011374106635,0.023634362654386513,0.15373471518946694,0.07491210201539855,0.9701505690514213,0.028048782169096887,0.0003,26.59440040588379,464.66706681251526,15.61366891860962
10
+ 9,0.0244467155419686,0.15635445481970955,0.07858170639554084,0.9691203908836945,0.020375159266375113,0.14274158212089116,0.06738522627825898,0.974266836873002,0.026637090500559478,0.0003,30.914405822753906,446.49894666671753,15.61366891860962
11
+ 10,0.021730341876117505,0.14741214968962874,0.07422529699942916,0.9725515493643391,0.024054128541071057,0.1550939345721523,0.07450153486611187,0.9696204183961069,0.026002763604478115,0.0003,10.194475173950195,451.94559621810913,15.61366891860962
12
+ 11,0.021080626250557278,0.14519168795271056,0.07274680407243074,0.9733722308184185,0.022258744493567427,0.14919364763141701,0.07787581567146928,0.9718879300246548,0.025471897251637635,0.0003,38.406463623046875,448.4989433288574,15.61366891860962
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/FNO-Large/E-AC80/checkpoints/FNO-Large_E-AC80_best.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f7783fa4d617a8778b005e7c8666113ef8c08edd017d7cc94080bbc38abff3e5
3
+ size 203221651
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/FNO-Large/E-AC80/config.json ADDED
@@ -0,0 +1,87 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "FNO-Large",
3
+ "run_id": "E-AC80",
4
+ "axis": "ac",
5
+ "seed": 42,
6
+ "epochs": 50,
7
+ "batch_size": 64,
8
+ "eval_batch_size": 64,
9
+ "lr": 0.001,
10
+ "base_lr": 0.001,
11
+ "bilstm_lr": 0.0003,
12
+ "transformer_lr": 0.0003,
13
+ "weight_decay": 0.0001,
14
+ "scheduler_step": 20,
15
+ "scheduler_gamma": 0.5,
16
+ "grad_clip": 0.0,
17
+ "train_gms_used": 3000,
18
+ "use_scales_count": 46,
19
+ "train_gms": 2400,
20
+ "val_gms": 600,
21
+ "test_gms": 474,
22
+ "train_samples": 110400,
23
+ "val_samples": 27600,
24
+ "test_samples": 27018,
25
+ "scale_indices": [
26
+ 0,
27
+ 1,
28
+ 2,
29
+ 4,
30
+ 5,
31
+ 6,
32
+ 7,
33
+ 9,
34
+ 10,
35
+ 11,
36
+ 12,
37
+ 14,
38
+ 15,
39
+ 16,
40
+ 17,
41
+ 19,
42
+ 20,
43
+ 21,
44
+ 22,
45
+ 24,
46
+ 25,
47
+ 26,
48
+ 27,
49
+ 29,
50
+ 30,
51
+ 31,
52
+ 32,
53
+ 34,
54
+ 35,
55
+ 36,
56
+ 37,
57
+ 39,
58
+ 40,
59
+ 41,
60
+ 42,
61
+ 44,
62
+ 45,
63
+ 46,
64
+ 47,
65
+ 49,
66
+ 50,
67
+ 51,
68
+ 52,
69
+ 54,
70
+ 55,
71
+ 56
72
+ ],
73
+ "split": "fixed_474_test_full_57_factors_reduced_train_val",
74
+ "params": {
75
+ "numel": 8505793,
76
+ "numel_m": 8.505793,
77
+ "real_params": 16927169,
78
+ "real_params_m": 16.927169,
79
+ "trainable": 8505793
80
+ },
81
+ "architecture": {
82
+ "n_modes": 512,
83
+ "hidden_channels": 64,
84
+ "n_layers": 8,
85
+ "domain_padding": 0.1
86
+ }
87
+ }
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/FNO-Large/E-AC80/dataset_indices.pkl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:350b63e638363a4dd6ba42f937de91a93b2ca9eac0a362d75e26696effba9c3e
3
+ size 1348786
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/FNO-Large/E-AC80/split_info.txt ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ Canonical split: fixed 474-GM test + 3000-GM pool + 80/20 train/val
2
+ Train GMs: 2400
3
+ Val GMs: 600
4
+ Test GMs: 474
5
+ Train samples: 110400
6
+ Val samples: 27600
7
+ Test samples: 27018
8
+ Scale count: 46
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/artifact/runs/FNO-Large/E-AC80/training_log.csv ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,train_mse,train_rmse,train_mae,train_r2,val_mse,val_rmse,val_mae,val_r2,val_pfa_rmse_g,lr,train_max_grad_norm,epoch_s,gpu_peak_gb
2
+ 1,0.12126523962638229,0.3482315890702368,0.1914170190193803,0.8468251003712787,0.0854544422359743,0.29232591783140666,0.15890443194255738,0.8920738202221583,0.048487354404914654,0.001,0.0,128.10569858551025,3.2975897789001465
3
+ 2,0.05929946157114517,0.2435148077040597,0.13610364469518985,0.9250965148746828,0.07312266810615281,0.27041203395217606,0.1437314337659458,0.9076484496608438,0.047816688268088825,0.001,0.0,128.9756669998169,3.2975897789001465
4
+ 3,0.04637638465918205,0.2153517695752279,0.12109787283339939,0.9414201620815403,0.06454020317925348,0.2540476395860695,0.13580759872952522,0.9184878235822105,0.038498340768282686,0.001,0.0,135.930326461792,3.2975897789001465
5
+ 4,0.04000723249057641,0.20001808040918803,0.11257282902961768,0.9494652891083006,0.058650483595990906,0.24217861919663947,0.12833550762379228,0.9259263477589756,0.03806331958857807,0.001,0.0,136.83617639541626,3.2975897789001465
6
+ 5,0.036330326692148106,0.19060515914357645,0.10671169070036515,0.9541097336232598,0.05497571432657288,0.23446900504453222,0.12319825240130586,0.9305674617658767,0.039150555424181686,0.001,0.0,130.77259373664856,3.2975897789001465
7
+ 6,0.032304933957454665,0.17973573366878012,0.10070723773163874,0.9591943657915014,0.054310605479069955,0.23304635907705135,0.12282209427838164,0.9314074725970188,0.035785377843268754,0.001,0.0,128.2022123336792,3.2975897789001465
8
+ 7,0.03049295019029995,0.1746223072528248,0.09767157722510002,0.9614831537277754,0.055501338088178406,0.23558721970467414,0.12565363468778307,0.9299036160592498,0.03600722001845966,0.001,0.0,129.8162488937378,3.2975897789001465
9
+ 8,0.02909736251093915,0.1705794902997988,0.09518576121159798,0.9632459755354097,0.05066162209036269,0.225081367710352,0.11636871837966108,0.9360160198756785,0.037982136847715786,0.001,0.0,129.5686159133911,3.2975897789001465
10
+ 9,0.027163695705985102,0.16481412471625453,0.09192266679109583,0.9656884662741119,0.05045014178437311,0.22461109007431737,0.11590322549866018,0.9362831126203796,0.03537878408340991,0.001,0.0,128.9032428264618,3.2975897789001465
11
+ 10,0.02621401257869702,0.16190741977654088,0.09032827077082389,0.9668880484070864,0.04989621667981724,0.2233746106427882,0.11719407527315444,0.9369827020022009,0.03562765926692314,0.001,0.0,129.40835976600647,3.2975897789001465
12
+ 11,0.025145819435303914,0.15857433410014343,0.08836233472741169,0.9682373252403534,0.05087969314059198,0.22556527467806736,0.1152013996924064,0.9357406032354716,0.0374229647667399,0.001,0.0,128.16395592689514,3.2975897789001465
13
+ 12,0.024458416406014115,0.15639186809426542,0.08702623664303102,0.9691056110799438,0.052032257801000625,0.22810580396167177,0.11768575493706597,0.9342849515748902,0.0388767406796468,0.001,0.0,128.4864103794098,3.2975897789001465
14
+ 13,0.023642742646443095,0.15376196749015375,0.08561581220986186,0.9701359206934714,0.04826372547389229,0.21969006685303796,0.11308186000952974,0.9390444852725179,0.035894282564442465,0.001,0.0,128.12085700035095,3.2975897789001465
15
+ 14,0.023520495371058367,0.15336393112807967,0.0853710681063887,0.970290336015536,0.04862757806897739,0.22051661631037556,0.11345924066368508,0.9385849512850237,0.03564173769720266,0.001,0.0,128.26137971878052,3.2975897789001465
16
+ 15,0.022678537796683933,0.15059395006667411,0.08371598903785005,0.9713538458043756,0.04952688244363536,0.22254636021205865,0.11433762654032684,0.9374491591240227,0.03980962786996683,0.001,0.0,128.1047179698944,3.2975897789001465
17
+ 16,0.02203259746975369,0.1484338151155379,0.08252155849880642,0.9721697584699589,0.049063159396406535,0.22150205280404633,0.11585750082559632,0.938034826242698,0.040955268108016564,0.001,0.0,128.97408151626587,3.2975897789001465
18
+ 17,0.02145742375028306,0.14648352723184632,0.08144568344042497,0.972896283045949,0.04715907270643446,0.2171613978275938,0.11005421203318426,0.9404396257713995,0.03724157202995056,0.001,0.0,158.32637453079224,3.2975897789001465
19
+ 18,0.021116587772830098,0.14531547671473297,0.08082675813684141,0.9733268063835854,0.04831047703987158,0.21979644455693906,0.11208858404334616,0.9389854395660251,0.0357738949179065,0.001,0.0,155.96678972244263,3.2975897789001465
20
+ 19,0.020791652871836788,0.14419310965450738,0.07998325807433197,0.9737372445097567,0.04770497151840136,0.21841467789139393,0.11340377671393795,0.939750173335935,0.03418194745102355,0.001,0.0,155.23586058616638,3.2975897789001465
21
+ 20,0.020154148002292798,0.14196530562885004,0.07888326609441047,0.9745425019637557,0.047837096218532986,0.2187169317143348,0.11235625350288724,0.9395833041391282,0.034959586950222186,0.0005,0.0,155.77941370010376,3.2975897789001465
22
+ 21,0.016622371821380467,0.12892777753990978,0.07188835752625396,0.9790036275240601,0.046131757734363205,0.21478304806097526,0.10750927389983393,0.9417370911513508,0.034446966298407974,0.0005,0.0,151.06660437583923,3.2975897789001465
23
+ 22,0.016007908285021206,0.1265223627862727,0.07109041942946577,0.9797797806441813,0.04755992997616385,0.21808239263215143,0.10904149977826842,0.9399333560839958,0.033406798239045,0.0005,0.0,152.66288781166077,3.2975897789001465
24
+ 23,0.015774981279972095,0.12559849234752818,0.07047280249259322,0.9800739998992536,0.04700343813246575,0.21680276320302228,0.10849588115894852,0.9406361871738297,0.033191314566666,0.0005,0.0,142.32469415664673,3.2975897789001465
25
+ 24,0.015628909617253547,0.1250156374908897,0.07020520279299233,0.9802585088278105,0.04820628689530967,0.2195593015458686,0.11018567020729544,0.9391170283281636,0.0353318776050214,0.0005,0.0,153.8207230567932,3.2975897789001465
26
+ 25,0.015429300189548068,0.12421473418861415,0.06974939872409987,0.9805106433149272,0.04629838561274579,0.21517059653388004,0.10704253331281137,0.9415266455631782,0.034112858224289014,0.0005,0.0,144.97609448432922,3.2975897789001465
27
+ 26,0.015186525836520725,0.12323362299519042,0.06916107382659174,0.9808173011585575,0.047017483392245527,0.21683515257505073,0.10975860021323973,0.9406184484677735,0.0360403468280849,0.0005,0.0,130.49029397964478,3.2975897789001465
28
+ 27,0.015193380833132831,0.12326143286986742,0.06914121773097825,0.980808642339524,0.047116979406974166,0.21706445910598576,0.10791809329120433,0.9404927882388632,0.034035001648779976,0.0005,0.0,132.4570004940033,3.2975897789001465
29
+ 28,0.014857319005698973,0.12189060261438932,0.0684557827979931,0.9812331352760246,0.04699193381784043,0.21677622982661274,0.10742088677890059,0.9406507167488405,0.03423982783743861,0.0005,0.0,132.6577525138855,3.2975897789001465
30
+ 29,0.014600823637736592,0.12083386792508378,0.06789330742435179,0.9815571246750994,0.046547075672978944,0.21574771301911624,0.10718745233360696,0.9412125581097072,0.03320874777015247,0.0005,0.0,136.3624129295349,3.2975897789001465
31
+ 30,0.014563398434920011,0.12067890633793468,0.0678096621247075,0.981604397963712,0.04784036488408628,0.2187244039518368,0.1087651472333088,0.9395791759208997,0.03324216853782224,0.0005,0.0,135.98989081382751,3.2975897789001465
32
+ 31,0.014507189605270606,0.12044579529925736,0.06761308895092656,0.9816753974842969,0.04632100686335909,0.215223155964592,0.10674935915444785,0.94149807565977,0.03476264808642855,0.0005,0.0,136.51144003868103,3.2975897789001465
33
+ 32,0.014402713177077437,0.12001130437203587,0.06748463670813519,0.9818073658001966,0.04738583474274419,0.21768287654922283,0.10937875403362772,0.9401532326985856,0.03607037970519716,0.0005,0.0,136.3166687488556,3.2975897789001465
34
+ 33,0.01414743930687651,0.11894300865068325,0.06691779950275513,0.9821298122619507,0.04690096268216193,0.21656630089227163,0.10751406622697765,0.940765610332918,0.03261743865530526,0.0005,0.0,136.43114948272705,3.2975897789001465
35
+ 34,0.014076243191410378,0.11864334448847258,0.06678113424844788,0.9822197428707383,0.048017722175340144,0.21912946441622164,0.10975592298129906,0.9393551794334156,0.034261920120027656,0.0005,0.0,136.35562896728516,3.2975897789001465
36
+ 35,0.014130862094307868,0.1188733026979055,0.06684003050873245,0.9821507516222943,0.04809069939065095,0.21929591740534285,0.10998940264568237,0.9392630115852237,0.03265582847987373,0.0005,0.0,135.19205689430237,3.2975897789001465
37
+ 36,0.013998222013095727,0.11831408205744456,0.0665592079317397,0.9823182945358663,0.047685280166109976,0.21836959533348496,0.10897006052210712,0.9397750428731043,0.033574955239237016,0.0005,0.0,135.7451856136322,3.2975897789001465
38
+ 37,0.01380056140328375,0.11747579071146425,0.06598070326358224,0.9825679674348075,0.04638696492807877,0.21537633325896968,0.10802149682381303,0.9414147727703651,0.03400947052149165,0.0005,0.0,136.84096097946167,3.2975897789001465
experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80/manifest.json ADDED
@@ -0,0 +1,181 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "objective": "O2_data_efficiency",
3
+ "run_id": "2026-06-09_xgb2_partial_EAC80",
4
+ "server": "XG-Boost-2",
5
+ "repo": "HKUJasonJiang/Seismic-FNO",
6
+ "branch": "sync/xgb2-o1-o3-o2partial",
7
+ "commit": "df66472",
8
+ "hf_repo": "JasonXF/SeFNO",
9
+ "path_in_repo": "experiments/O2_data_efficiency/2026-06-09_xgb2_partial_EAC80",
10
+ "status": "partial",
11
+ "created_at": "2026-06-09T06:55:31Z",
12
+ "notes": "Partial O2 data-efficiency artifact from XG2. Reduced runs pending; observed package includes E-AC80 and baseline rows.",
13
+ "artifacts": [
14
+ {
15
+ "path": "artifact/compiled/o2_data_efficiency_metrics.csv",
16
+ "size_bytes": 1851,
17
+ "sha256": "349165424a741cb17b99ad417241aff09f9be042b3bb201a2bf7ae169a4ac166"
18
+ },
19
+ {
20
+ "path": "artifact/compiled/o2_data_efficiency_metrics.json",
21
+ "size_bytes": 3195,
22
+ "sha256": "22b1b71e424ec9d7f5625199dae6a86e6533cb45d3dd2673dc1bdf78256958f9"
23
+ },
24
+ {
25
+ "path": "artifact/logs/o2_gpu0_bilstm.log",
26
+ "size_bytes": 1029,
27
+ "sha256": "48e2e2e99566e78e3216cc579f2ff73c91a70bdc668340b0c56e1bb7d8503322"
28
+ },
29
+ {
30
+ "path": "artifact/logs/o2_gpu1_fno_transformer.log",
31
+ "size_bytes": 3446,
32
+ "sha256": "2fb820af788c7cb9bcbcfc7b32dfa33b7b7a4f283d897543542fa310e7383ced"
33
+ },
34
+ {
35
+ "path": "artifact/logs/o2_postprocess_after_training.log",
36
+ "size_bytes": 1312,
37
+ "sha256": "e5cc3965b73a19243cd078e7dff65c02da6e981a82ffecf3948c62fe42a764ad"
38
+ },
39
+ {
40
+ "path": "artifact/report/O2_data_efficiency_report.md",
41
+ "size_bytes": 1190,
42
+ "sha256": "512573604977122c0dac4001e8f28c22c555286647cb47bc8e1d80722085e8f5"
43
+ },
44
+ {
45
+ "path": "artifact/report/O2_data_efficiency_report.pdf",
46
+ "size_bytes": 156136,
47
+ "sha256": "da7a095d543ae6769b578905ba8503063d747dc8706c22fafab2a4d6e907558b"
48
+ },
49
+ {
50
+ "path": "artifact/report/O2_data_efficiency_report.tex",
51
+ "size_bytes": 1794,
52
+ "sha256": "411d83f9a63c8ffadf02320de2bc6117c7ed5116e2f0604665740cdcb668ec4b"
53
+ },
54
+ {
55
+ "path": "artifact/report/figures/fig1_mse_degradation.pdf",
56
+ "size_bytes": 16510,
57
+ "sha256": "1ecad6bc066d1edd1c18bad7cd8a528d8a0831840cee7d2e7a04582293756701"
58
+ },
59
+ {
60
+ "path": "artifact/report/figures/fig1_mse_degradation.png",
61
+ "size_bytes": 57425,
62
+ "sha256": "5103ef2df5f0f77dc48a71fc75a48dc6f199b89e518e4f443f46e6fffe32eb45"
63
+ },
64
+ {
65
+ "path": "artifact/report/figures/fig2_r2_retention.pdf",
66
+ "size_bytes": 17662,
67
+ "sha256": "97ec63ff5bcd1043cf70b740e494e5d747c6d38e1d75847bc5eb160f8235fceb"
68
+ },
69
+ {
70
+ "path": "artifact/report/figures/fig2_r2_retention.png",
71
+ "size_bytes": 68361,
72
+ "sha256": "bb7487f5f98e87e2018b478bfada26e259f44bc9064b85545db796b826d2ffa7"
73
+ },
74
+ {
75
+ "path": "artifact/report/figures/fig3_retention_heatmap.pdf",
76
+ "size_bytes": 22018,
77
+ "sha256": "ccc830f5d391373474397c19c9df7bd8e6195eb52e3af7334f2d12406a935dcd"
78
+ },
79
+ {
80
+ "path": "artifact/report/figures/fig3_retention_heatmap.png",
81
+ "size_bytes": 95961,
82
+ "sha256": "b5766f4e0b16ea2bbacf6029bf643c834da0d95138b9c5c9c7f7af56b13ac2e7"
83
+ },
84
+ {
85
+ "path": "artifact/report/figures/fig4_training_curves.pdf",
86
+ "size_bytes": 15689,
87
+ "sha256": "c48e3f35de8d6493cb6c34d3b9f95c2cbdac5ecf146756d074063924d6b2dcb6"
88
+ },
89
+ {
90
+ "path": "artifact/report/figures/fig4_training_curves.png",
91
+ "size_bytes": 63726,
92
+ "sha256": "7e1b2bf8366b2ba308aa2a31f31e172ec2d43447563eed2728bcf8f9193d3e0f"
93
+ },
94
+ {
95
+ "path": "artifact/report/figures/fig5_results_table.pdf",
96
+ "size_bytes": 20108,
97
+ "sha256": "09bfd5a7bf4337f9a61caf5491c799806a98564a45e66536abf41c08b5136f49"
98
+ },
99
+ {
100
+ "path": "artifact/report/figures/fig5_results_table.png",
101
+ "size_bytes": 60145,
102
+ "sha256": "a5636233670f7a5ba05aa6e452c617c833a1b63e40189d759670e8e22c151f49"
103
+ },
104
+ {
105
+ "path": "artifact/report/o2_report_conclusions.json",
106
+ "size_bytes": 205,
107
+ "sha256": "ca751876fee6278396325b812af266d1a7d36d8bf28d70f3674937ca961931f1"
108
+ },
109
+ {
110
+ "path": "artifact/report/o2_report_summary.csv",
111
+ "size_bytes": 1849,
112
+ "sha256": "7c55b20f40c2612230f082a288154a42a0fe32c590a0cb9e71baa20885123117"
113
+ },
114
+ {
115
+ "path": "artifact/run_scripts/run_gpu0_bilstm_o2.sh",
116
+ "size_bytes": 623,
117
+ "sha256": "a3f4d4536156591b59812ed94c76d7c21000b4c4808a26fbcc9da63e8a0a3331"
118
+ },
119
+ {
120
+ "path": "artifact/run_scripts/run_gpu1_fno_transformer_o2.sh",
121
+ "size_bytes": 638,
122
+ "sha256": "328b320a8982264c82e1099a652aec0256c2eecbe762a4be910a0a14fb80bf27"
123
+ },
124
+ {
125
+ "path": "artifact/run_scripts/run_o2_postprocess_after_training.sh",
126
+ "size_bytes": 990,
127
+ "sha256": "54d2481bd6c67af9467db7381e672b687c3ae72f543ecaad3165f531a8a1d41f"
128
+ },
129
+ {
130
+ "path": "artifact/runs/BiLSTM/E-AC80/checkpoints/BiLSTM_E-AC80_best.pt",
131
+ "size_bytes": 100941869,
132
+ "sha256": "0fe8a5bcecaaf3ad346a6974ba7419e084e48d480fb36579d56bc855df0435da"
133
+ },
134
+ {
135
+ "path": "artifact/runs/BiLSTM/E-AC80/config.json",
136
+ "size_bytes": 1103,
137
+ "sha256": "03b6252de77430eb6d8af99007af97863fbd748ae2f3d6ce7ee5e947211c75e8"
138
+ },
139
+ {
140
+ "path": "artifact/runs/BiLSTM/E-AC80/dataset_indices.pkl",
141
+ "size_bytes": 1348786,
142
+ "sha256": "350b63e638363a4dd6ba42f937de91a93b2ca9eac0a362d75e26696effba9c3e"
143
+ },
144
+ {
145
+ "path": "artifact/runs/BiLSTM/E-AC80/split_info.txt",
146
+ "size_bytes": 188,
147
+ "sha256": "7e8ec4a0b42d205afaf22b2349ef20e8cb97613c902ca4b400a28d376b58ab21"
148
+ },
149
+ {
150
+ "path": "artifact/runs/BiLSTM/E-AC80/training_log.csv",
151
+ "size_bytes": 2806,
152
+ "sha256": "c039a47a7076f6cfa7d2d9df1361e5f12f32e0c8ea90ada62ebcc1dc039835a8"
153
+ },
154
+ {
155
+ "path": "artifact/runs/FNO-Large/E-AC80/checkpoints/FNO-Large_E-AC80_best.pt",
156
+ "size_bytes": 203221651,
157
+ "sha256": "f7783fa4d617a8778b005e7c8666113ef8c08edd017d7cc94080bbc38abff3e5"
158
+ },
159
+ {
160
+ "path": "artifact/runs/FNO-Large/E-AC80/config.json",
161
+ "size_bytes": 1224,
162
+ "sha256": "4f81b49ff8545c2721603d28252713c9b087818c6ee04593bef2ea484e766e2b"
163
+ },
164
+ {
165
+ "path": "artifact/runs/FNO-Large/E-AC80/dataset_indices.pkl",
166
+ "size_bytes": 1348786,
167
+ "sha256": "350b63e638363a4dd6ba42f937de91a93b2ca9eac0a362d75e26696effba9c3e"
168
+ },
169
+ {
170
+ "path": "artifact/runs/FNO-Large/E-AC80/split_info.txt",
171
+ "size_bytes": 188,
172
+ "sha256": "7e8ec4a0b42d205afaf22b2349ef20e8cb97613c902ca4b400a28d376b58ab21"
173
+ },
174
+ {
175
+ "path": "artifact/runs/FNO-Large/E-AC80/training_log.csv",
176
+ "size_bytes": 8647,
177
+ "sha256": "cd5f871fade283fdc737d33196bd007af8ad424783c0348609d36c71bf18fd51"
178
+ }
179
+ ],
180
+ "coverage": "partial; E-AC80 package plus E-Base references; not final O2 matrix"
181
+ }