Swahili transformer model bake off
pretrained swahili transformer, lr=0.0001, pos tagging
xxxxxxxxxx
1
2022-06-11 17:40:51,753 - INFO - allennlp.common.util - Metrics: {
2
"best_epoch": 4,
3
"peak_worker_0_memory_MB": 5448.7890625,
4
"peak_gpu_0_memory_MB": 9332.94873046875,
5
"training_duration": "1:26:00.349669",
6
"epoch": 9,
7
"training_accuracy": 0.996923352964769,
8
"training_accuracy3": 0.9999548500435081,
9
"training_precision": 0.9857814311981201,
10
"training_recall": 0.9809743165969849,
11
"training_fscore": 0.9833330512046814,
12
"training_loss": 0.008818848443899869,
13
"training_worker_0_memory_MB": 5448.7890625,
14
"training_gpu_0_memory_MB": 9332.94873046875,
15
"validation_accuracy": 0.9915935820231985,
16
"validation_accuracy3": 0.9994341834054076,
17
"validation_precision": 0.9857875108718872,
18
"validation_recall": 0.9809861779212952,
19
"validation_fscore": 0.9833420515060425,
20
"validation_loss": 0.06279067079275392,
21
"best_validation_accuracy": 0.989734470355252,
22
"best_validation_accuracy3": 0.9997575071737461,
23
"best_validation_precision": 0.9777423143386841,
24
"best_validation_recall": 0.9695621728897095,
25
"best_validation_fscore": 0.9735202789306641,
26
"best_validation_loss": 0.05510239019787424,
27
"test_accuracy": 0.9954022988505747,
28
"test_accuracy3": 0.9996934865900383,
29
"test_precision": 0.9858008623123169,
30
"test_recall": 0.9810028076171875,
31
"test_fscore": 0.9833571910858154,
32
"test_loss": 0.0319001576157281
33
}
pretrained swahili transformer, lr=0.00001, pos tagging
xxxxxxxxxx
1
"best_epoch": 7,
2
"peak_worker_0_memory_MB": 5449.3203125,
3
"peak_gpu_0_memory_MB": 9332.94873046875,
4
"training_duration": "1:25:54.586565",
5
"epoch": 9,
6
"training_accuracy": 0.9972294344879995,
7
"training_accuracy3": 0.9999730272987192,
8
"training_precision": 0.9764115214347839,
9
"training_recall": 0.9566909074783325,
10
"training_fscore": 0.9659566879272461,
11
"training_loss": 0.007170079944642032,
12
"training_worker_0_memory_MB": 5449.3203125,
13
"training_gpu_0_memory_MB": 9332.94873046875,
14
"validation_accuracy": 0.9921189831467486,
15
"validation_accuracy3": 0.9996766762316615,
16
"validation_precision": 0.9764279127120972,
17
"validation_recall": 0.9567330479621887,
18
"validation_fscore": 0.9659876227378845,
19
"validation_loss": 0.0693623513392432,
20
"best_validation_accuracy": 0.991916905791537,
21
"best_validation_accuracy3": 0.9996362607606192,
22
"best_validation_precision": 0.9715831279754639,
23
"best_validation_recall": 0.9473865032196045,
24
"best_validation_fscore": 0.9585748910903931,
25
"best_validation_loss": 0.05559036874520399,
26
"test_accuracy": 0.9962835249042146,
27
"test_accuracy3": 0.9997318007662835,
28
"test_precision": 0.9764541387557983,
29
"test_recall": 0.9567817449569702,
30
"test_fscore": 0.9660264253616333,
31
"test_loss": 0.032172908316944124
32
}
pretrained swahili transformer, lr=0.0001, crf, pos tagging
xxxxxxxxxx
1
2022-06-11 22:44:50,984 - INFO - allennlp.common.util - Metrics: {
2
"best_epoch": 4,
3
"peak_worker_0_memory_MB": 5698.4609375,
4
"peak_gpu_0_memory_MB": 9332.60205078125,
5
"training_duration": "2:12:03.720062",
6
"epoch": 9,
7
"training_accuracy": 0.9975718705216521,
8
"training_accuracy3": 0.9978339748145334,
9
"training_loss": 1.5775271221179867,
10
"training_worker_0_memory_MB": 5698.4609375,
11
"training_gpu_0_memory_MB": 9332.60205078125,
12
"validation_accuracy": 0.9919573212625793,
13
"validation_accuracy3": 0.9926039687992564,
14
"validation_loss": 8.17987885513926,
15
"best_validation_accuracy": 0.9915127510811138,
16
"best_validation_accuracy3": 0.9925231378571717,
17
"best_validation_loss": 7.377661480167048,
18
"test_accuracy": 0.9957471264367816,
19
"test_accuracy3": 0.9962835249042146,
20
"test_loss": 3.963569987903942
21
}
pretrained swahili transformer, lr=0.0001, helsinki pos tagging
xxxxxxxxxx
1
{
2
"best_epoch": 5,
3
"peak_worker_0_memory_MB": 5583.07421875,
4
"peak_gpu_0_memory_MB": 9335.52783203125,
5
"training_duration": "1:24:44.563542",
6
"epoch": 9,
7
"training_accuracy": 0.9965898688553921,
8
"training_accuracy3": 0.9999364222373633,
9
"training_precision": 0.9467342495918274,
10
"training_recall": 0.8755066990852356,
11
"training_fscore": 0.8997867107391357,
12
"training_loss": 0.01112793169544195,
13
"training_worker_0_memory_MB": 5583.07421875,
14
"training_gpu_0_memory_MB": 9335.52783203125,
15
"validation_accuracy": 0.9915556244189649,
16
"validation_accuracy3": 0.999535171986365,
17
"validation_precision": 0.9467248916625977,
18
"validation_recall": 0.8757268190383911,
19
"validation_fscore": 0.8999462723731995,
20
"validation_loss": 0.06682281847298145,
21
"best_validation_accuracy": 0.9905484970560893,
22
"best_validation_accuracy3": 0.9994964363185621,
23
"best_validation_precision": 0.9082106351852417,
24
"best_validation_recall": 0.8251690864562988,
25
"best_validation_fscore": 0.8477039337158203,
26
"best_validation_loss": 0.05897250342555344,
27
"test_accuracy": 0.9952177563611142,
28
"test_accuracy3": 0.9998174716168364,
29
"test_precision": 0.9467272162437439,
30
"test_recall": 0.8757625222206116,
31
"test_fscore": 0.8999730348587036,
32
"test_loss": 0.02968891077139415
33
}
No Comments