mt5-small-gos-nld-nonpretrained / trainer_state.json
j-arkes's picture
first version
c7b4702
{
"best_metric": 1.4148167371749878,
"best_model_checkpoint": "/scratch/s3545881/dumped/translation/mt5/3734190/checkpoint-4600",
"epoch": 18.231144872490503,
"global_step": 5600,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.03,
"learning_rate": 2.5e-06,
"loss": 16.635,
"step": 10
},
{
"epoch": 0.07,
"learning_rate": 5e-06,
"loss": 16.8135,
"step": 20
},
{
"epoch": 0.1,
"learning_rate": 7.5e-06,
"loss": 16.2077,
"step": 30
},
{
"epoch": 0.13,
"learning_rate": 1e-05,
"loss": 15.7712,
"step": 40
},
{
"epoch": 0.16,
"learning_rate": 1.25e-05,
"loss": 14.7744,
"step": 50
},
{
"epoch": 0.2,
"learning_rate": 1.5e-05,
"loss": 14.0924,
"step": 60
},
{
"epoch": 0.23,
"learning_rate": 1.7500000000000002e-05,
"loss": 13.2153,
"step": 70
},
{
"epoch": 0.26,
"learning_rate": 2e-05,
"loss": 12.5009,
"step": 80
},
{
"epoch": 0.29,
"learning_rate": 2.2499999999999998e-05,
"loss": 11.6579,
"step": 90
},
{
"epoch": 0.33,
"learning_rate": 2.5e-05,
"loss": 10.6585,
"step": 100
},
{
"epoch": 0.36,
"learning_rate": 2.75e-05,
"loss": 9.6627,
"step": 110
},
{
"epoch": 0.39,
"learning_rate": 3e-05,
"loss": 9.1927,
"step": 120
},
{
"epoch": 0.42,
"learning_rate": 3.2500000000000004e-05,
"loss": 8.4296,
"step": 130
},
{
"epoch": 0.46,
"learning_rate": 3.5000000000000004e-05,
"loss": 7.9397,
"step": 140
},
{
"epoch": 0.49,
"learning_rate": 3.75e-05,
"loss": 7.5002,
"step": 150
},
{
"epoch": 0.52,
"learning_rate": 4e-05,
"loss": 7.0869,
"step": 160
},
{
"epoch": 0.55,
"learning_rate": 4.25e-05,
"loss": 6.8123,
"step": 170
},
{
"epoch": 0.59,
"learning_rate": 4.4999999999999996e-05,
"loss": 6.4624,
"step": 180
},
{
"epoch": 0.62,
"learning_rate": 4.75e-05,
"loss": 6.2002,
"step": 190
},
{
"epoch": 0.65,
"learning_rate": 5e-05,
"loss": 5.9634,
"step": 200
},
{
"epoch": 0.65,
"eval_bp": 1.0,
"eval_counts": [
8158,
3455,
1823,
1029
],
"eval_loss": 3.1029841899871826,
"eval_precisions": [
2.863641788529988,
1.2277067291121069,
0.6539183086365283,
0.3724954297815345
],
"eval_ref_len": 41379,
"eval_runtime": 262.8531,
"eval_samples_per_second": 13.175,
"eval_score": 0.9619764567851472,
"eval_steps_per_second": 13.175,
"eval_sys_len": 284882,
"eval_totals": [
284882,
281419,
278781,
276245
],
"step": 200
},
{
"epoch": 0.68,
"learning_rate": 5.25e-05,
"loss": 5.8077,
"step": 210
},
{
"epoch": 0.72,
"learning_rate": 5.5e-05,
"loss": 5.693,
"step": 220
},
{
"epoch": 0.75,
"learning_rate": 5.75e-05,
"loss": 5.5207,
"step": 230
},
{
"epoch": 0.78,
"learning_rate": 6e-05,
"loss": 5.4243,
"step": 240
},
{
"epoch": 0.81,
"learning_rate": 6.25e-05,
"loss": 5.3032,
"step": 250
},
{
"epoch": 0.85,
"learning_rate": 6.500000000000001e-05,
"loss": 5.2149,
"step": 260
},
{
"epoch": 0.88,
"learning_rate": 6.75e-05,
"loss": 5.1234,
"step": 270
},
{
"epoch": 0.91,
"learning_rate": 7.000000000000001e-05,
"loss": 5.0828,
"step": 280
},
{
"epoch": 0.94,
"learning_rate": 7.25e-05,
"loss": 4.9444,
"step": 290
},
{
"epoch": 0.98,
"learning_rate": 7.5e-05,
"loss": 4.8644,
"step": 300
},
{
"epoch": 1.01,
"learning_rate": 7.75e-05,
"loss": 4.8276,
"step": 310
},
{
"epoch": 1.04,
"learning_rate": 8e-05,
"loss": 4.8093,
"step": 320
},
{
"epoch": 1.07,
"learning_rate": 8.25e-05,
"loss": 4.7373,
"step": 330
},
{
"epoch": 1.11,
"learning_rate": 8.5e-05,
"loss": 4.6358,
"step": 340
},
{
"epoch": 1.14,
"learning_rate": 8.75e-05,
"loss": 4.6567,
"step": 350
},
{
"epoch": 1.17,
"learning_rate": 8.999999999999999e-05,
"loss": 4.4909,
"step": 360
},
{
"epoch": 1.2,
"learning_rate": 9.25e-05,
"loss": 4.5973,
"step": 370
},
{
"epoch": 1.24,
"learning_rate": 9.5e-05,
"loss": 4.4554,
"step": 380
},
{
"epoch": 1.27,
"learning_rate": 9.750000000000001e-05,
"loss": 4.5076,
"step": 390
},
{
"epoch": 1.3,
"learning_rate": 0.0001,
"loss": 4.4274,
"step": 400
},
{
"epoch": 1.3,
"eval_bp": 1.0,
"eval_counts": [
20993,
12057,
7745,
5231
],
"eval_loss": 2.268582582473755,
"eval_precisions": [
27.83626816590644,
16.75677178157964,
11.30788997255154,
8.037923139568832
],
"eval_ref_len": 41379,
"eval_runtime": 232.9413,
"eval_samples_per_second": 14.866,
"eval_score": 14.349332444590503,
"eval_steps_per_second": 14.866,
"eval_sys_len": 75416,
"eval_totals": [
75416,
71953,
68492,
65079
],
"step": 400
},
{
"epoch": 1.33,
"learning_rate": 0.0001025,
"loss": 4.4082,
"step": 410
},
{
"epoch": 1.37,
"learning_rate": 0.000105,
"loss": 4.3035,
"step": 420
},
{
"epoch": 1.4,
"learning_rate": 0.0001075,
"loss": 4.3256,
"step": 430
},
{
"epoch": 1.43,
"learning_rate": 0.00011,
"loss": 4.3323,
"step": 440
},
{
"epoch": 1.47,
"learning_rate": 0.00011250000000000001,
"loss": 4.2534,
"step": 450
},
{
"epoch": 1.5,
"learning_rate": 0.000115,
"loss": 4.291,
"step": 460
},
{
"epoch": 1.53,
"learning_rate": 0.0001175,
"loss": 4.2201,
"step": 470
},
{
"epoch": 1.56,
"learning_rate": 0.00012,
"loss": 4.3005,
"step": 480
},
{
"epoch": 1.6,
"learning_rate": 0.0001225,
"loss": 4.2619,
"step": 490
},
{
"epoch": 1.63,
"learning_rate": 0.000125,
"loss": 4.1464,
"step": 500
},
{
"epoch": 1.66,
"learning_rate": 0.0001275,
"loss": 4.104,
"step": 510
},
{
"epoch": 1.69,
"learning_rate": 0.00013000000000000002,
"loss": 4.1574,
"step": 520
},
{
"epoch": 1.73,
"learning_rate": 0.00013250000000000002,
"loss": 4.1,
"step": 530
},
{
"epoch": 1.76,
"learning_rate": 0.000135,
"loss": 4.078,
"step": 540
},
{
"epoch": 1.79,
"learning_rate": 0.0001375,
"loss": 4.1266,
"step": 550
},
{
"epoch": 1.82,
"learning_rate": 0.00014000000000000001,
"loss": 3.9953,
"step": 560
},
{
"epoch": 1.86,
"learning_rate": 0.0001425,
"loss": 4.0587,
"step": 570
},
{
"epoch": 1.89,
"learning_rate": 0.000145,
"loss": 4.0022,
"step": 580
},
{
"epoch": 1.92,
"learning_rate": 0.0001475,
"loss": 3.963,
"step": 590
},
{
"epoch": 1.95,
"learning_rate": 0.00015,
"loss": 4.0209,
"step": 600
},
{
"epoch": 1.95,
"eval_bp": 1.0,
"eval_counts": [
22518,
13520,
9025,
6338
],
"eval_loss": 2.0687694549560547,
"eval_precisions": [
44.010554089709764,
28.342627143515994,
20.398246089865292,
15.523280021553308
],
"eval_ref_len": 41379,
"eval_runtime": 179.1653,
"eval_samples_per_second": 19.329,
"eval_score": 25.06936390823504,
"eval_steps_per_second": 19.329,
"eval_sys_len": 51165,
"eval_totals": [
51165,
47702,
44244,
40829
],
"step": 600
},
{
"epoch": 1.99,
"learning_rate": 0.0001525,
"loss": 3.922,
"step": 610
},
{
"epoch": 2.02,
"learning_rate": 0.000155,
"loss": 3.872,
"step": 620
},
{
"epoch": 2.05,
"learning_rate": 0.0001575,
"loss": 3.9436,
"step": 630
},
{
"epoch": 2.08,
"learning_rate": 0.00016,
"loss": 3.8825,
"step": 640
},
{
"epoch": 2.12,
"learning_rate": 0.00016250000000000002,
"loss": 3.8445,
"step": 650
},
{
"epoch": 2.15,
"learning_rate": 0.000165,
"loss": 3.8438,
"step": 660
},
{
"epoch": 2.18,
"learning_rate": 0.0001675,
"loss": 3.9445,
"step": 670
},
{
"epoch": 2.21,
"learning_rate": 0.00017,
"loss": 3.8441,
"step": 680
},
{
"epoch": 2.25,
"learning_rate": 0.0001725,
"loss": 3.8509,
"step": 690
},
{
"epoch": 2.28,
"learning_rate": 0.000175,
"loss": 3.8228,
"step": 700
},
{
"epoch": 2.31,
"learning_rate": 0.0001775,
"loss": 3.7646,
"step": 710
},
{
"epoch": 2.34,
"learning_rate": 0.00017999999999999998,
"loss": 3.782,
"step": 720
},
{
"epoch": 2.38,
"learning_rate": 0.0001825,
"loss": 3.8468,
"step": 730
},
{
"epoch": 2.41,
"learning_rate": 0.000185,
"loss": 3.7436,
"step": 740
},
{
"epoch": 2.44,
"learning_rate": 0.0001875,
"loss": 3.7842,
"step": 750
},
{
"epoch": 2.47,
"learning_rate": 0.00019,
"loss": 3.7974,
"step": 760
},
{
"epoch": 2.51,
"learning_rate": 0.00019250000000000002,
"loss": 3.7282,
"step": 770
},
{
"epoch": 2.54,
"learning_rate": 0.00019500000000000002,
"loss": 3.7219,
"step": 780
},
{
"epoch": 2.57,
"learning_rate": 0.0001975,
"loss": 3.6277,
"step": 790
},
{
"epoch": 2.6,
"learning_rate": 0.0002,
"loss": 3.7015,
"step": 800
},
{
"epoch": 2.6,
"eval_bp": 1.0,
"eval_counts": [
24390,
15185,
10398,
7459
],
"eval_loss": 1.9690614938735962,
"eval_precisions": [
54.65668698458229,
36.89171788829231,
27.56116309274525,
21.719760060567236
],
"eval_ref_len": 41379,
"eval_runtime": 156.3814,
"eval_samples_per_second": 22.145,
"eval_score": 33.14600462748512,
"eval_steps_per_second": 22.145,
"eval_sys_len": 44624,
"eval_totals": [
44624,
41161,
37727,
34342
],
"step": 800
},
{
"epoch": 2.64,
"learning_rate": 0.00020250000000000002,
"loss": 3.794,
"step": 810
},
{
"epoch": 2.67,
"learning_rate": 0.000205,
"loss": 3.6069,
"step": 820
},
{
"epoch": 2.7,
"learning_rate": 0.0002075,
"loss": 3.7057,
"step": 830
},
{
"epoch": 2.73,
"learning_rate": 0.00021,
"loss": 3.6715,
"step": 840
},
{
"epoch": 2.77,
"learning_rate": 0.0002125,
"loss": 3.6437,
"step": 850
},
{
"epoch": 2.8,
"learning_rate": 0.000215,
"loss": 3.6129,
"step": 860
},
{
"epoch": 2.83,
"learning_rate": 0.0002175,
"loss": 3.6232,
"step": 870
},
{
"epoch": 2.86,
"learning_rate": 0.00022,
"loss": 3.5967,
"step": 880
},
{
"epoch": 2.9,
"learning_rate": 0.00022250000000000001,
"loss": 3.5749,
"step": 890
},
{
"epoch": 2.93,
"learning_rate": 0.00022500000000000002,
"loss": 3.6316,
"step": 900
},
{
"epoch": 2.96,
"learning_rate": 0.0002275,
"loss": 3.6169,
"step": 910
},
{
"epoch": 3.0,
"learning_rate": 0.00023,
"loss": 3.5349,
"step": 920
},
{
"epoch": 3.03,
"learning_rate": 0.0002325,
"loss": 3.4972,
"step": 930
},
{
"epoch": 3.06,
"learning_rate": 0.000235,
"loss": 3.534,
"step": 940
},
{
"epoch": 3.09,
"learning_rate": 0.0002375,
"loss": 3.5623,
"step": 950
},
{
"epoch": 3.13,
"learning_rate": 0.00024,
"loss": 3.4621,
"step": 960
},
{
"epoch": 3.16,
"learning_rate": 0.00024249999999999999,
"loss": 3.5827,
"step": 970
},
{
"epoch": 3.19,
"learning_rate": 0.000245,
"loss": 3.4537,
"step": 980
},
{
"epoch": 3.22,
"learning_rate": 0.0002475,
"loss": 3.504,
"step": 990
},
{
"epoch": 3.26,
"learning_rate": 0.00025,
"loss": 3.5469,
"step": 1000
},
{
"epoch": 3.26,
"eval_bp": 1.0,
"eval_counts": [
24938,
15828,
10952,
7917
],
"eval_loss": 1.8684660196304321,
"eval_precisions": [
52.90423861852433,
36.240412135088725,
27.213318424649025,
21.47740220281048
],
"eval_ref_len": 41379,
"eval_runtime": 163.2477,
"eval_samples_per_second": 21.213,
"eval_score": 32.53581668713074,
"eval_steps_per_second": 21.213,
"eval_sys_len": 47138,
"eval_totals": [
47138,
43675,
40245,
36862
],
"step": 1000
},
{
"epoch": 3.29,
"learning_rate": 0.0002525,
"loss": 3.4755,
"step": 1010
},
{
"epoch": 3.32,
"learning_rate": 0.000255,
"loss": 3.5018,
"step": 1020
},
{
"epoch": 3.35,
"learning_rate": 0.0002575,
"loss": 3.4461,
"step": 1030
},
{
"epoch": 3.39,
"learning_rate": 0.00026000000000000003,
"loss": 3.4587,
"step": 1040
},
{
"epoch": 3.42,
"learning_rate": 0.00026250000000000004,
"loss": 3.391,
"step": 1050
},
{
"epoch": 3.45,
"learning_rate": 0.00026500000000000004,
"loss": 3.4249,
"step": 1060
},
{
"epoch": 3.48,
"learning_rate": 0.0002675,
"loss": 3.5206,
"step": 1070
},
{
"epoch": 3.52,
"learning_rate": 0.00027,
"loss": 3.5261,
"step": 1080
},
{
"epoch": 3.55,
"learning_rate": 0.0002725,
"loss": 3.4603,
"step": 1090
},
{
"epoch": 3.58,
"learning_rate": 0.000275,
"loss": 3.4359,
"step": 1100
},
{
"epoch": 3.61,
"learning_rate": 0.0002775,
"loss": 3.4632,
"step": 1110
},
{
"epoch": 3.65,
"learning_rate": 0.00028000000000000003,
"loss": 3.4216,
"step": 1120
},
{
"epoch": 3.68,
"learning_rate": 0.0002825,
"loss": 3.3616,
"step": 1130
},
{
"epoch": 3.71,
"learning_rate": 0.000285,
"loss": 3.3384,
"step": 1140
},
{
"epoch": 3.74,
"learning_rate": 0.0002875,
"loss": 3.2927,
"step": 1150
},
{
"epoch": 3.78,
"learning_rate": 0.00029,
"loss": 3.4258,
"step": 1160
},
{
"epoch": 3.81,
"learning_rate": 0.0002925,
"loss": 3.4243,
"step": 1170
},
{
"epoch": 3.84,
"learning_rate": 0.000295,
"loss": 3.3272,
"step": 1180
},
{
"epoch": 3.87,
"learning_rate": 0.00029749999999999997,
"loss": 3.4568,
"step": 1190
},
{
"epoch": 3.91,
"learning_rate": 0.0003,
"loss": 3.4532,
"step": 1200
},
{
"epoch": 3.91,
"eval_bp": 1.0,
"eval_counts": [
25039,
16021,
11141,
8061
],
"eval_loss": 1.7631980180740356,
"eval_precisions": [
49.4558454640621,
33.96726455497604,
25.47329431132248,
19.976704996034893
],
"eval_ref_len": 41379,
"eval_runtime": 193.6778,
"eval_samples_per_second": 17.88,
"eval_score": 30.406875331612675,
"eval_steps_per_second": 17.88,
"eval_sys_len": 50629,
"eval_totals": [
50629,
47166,
43736,
40352
],
"step": 1200
},
{
"epoch": 3.94,
"learning_rate": 0.0003025,
"loss": 3.3871,
"step": 1210
},
{
"epoch": 3.97,
"learning_rate": 0.000305,
"loss": 3.3748,
"step": 1220
},
{
"epoch": 4.0,
"learning_rate": 0.0003075,
"loss": 3.3834,
"step": 1230
},
{
"epoch": 4.04,
"learning_rate": 0.00031,
"loss": 3.2802,
"step": 1240
},
{
"epoch": 4.07,
"learning_rate": 0.0003125,
"loss": 3.3111,
"step": 1250
},
{
"epoch": 4.1,
"learning_rate": 0.000315,
"loss": 3.2518,
"step": 1260
},
{
"epoch": 4.13,
"learning_rate": 0.0003175,
"loss": 3.3339,
"step": 1270
},
{
"epoch": 4.17,
"learning_rate": 0.00032,
"loss": 3.2785,
"step": 1280
},
{
"epoch": 4.2,
"learning_rate": 0.00032250000000000003,
"loss": 3.2316,
"step": 1290
},
{
"epoch": 4.23,
"learning_rate": 0.00032500000000000004,
"loss": 3.3004,
"step": 1300
},
{
"epoch": 4.26,
"learning_rate": 0.00032750000000000005,
"loss": 3.236,
"step": 1310
},
{
"epoch": 4.3,
"learning_rate": 0.00033,
"loss": 3.3492,
"step": 1320
},
{
"epoch": 4.33,
"learning_rate": 0.0003325,
"loss": 3.2794,
"step": 1330
},
{
"epoch": 4.36,
"learning_rate": 0.000335,
"loss": 3.1176,
"step": 1340
},
{
"epoch": 4.4,
"learning_rate": 0.0003375,
"loss": 3.2454,
"step": 1350
},
{
"epoch": 4.43,
"learning_rate": 0.00034,
"loss": 3.2039,
"step": 1360
},
{
"epoch": 4.46,
"learning_rate": 0.00034250000000000003,
"loss": 3.2624,
"step": 1370
},
{
"epoch": 4.49,
"learning_rate": 0.000345,
"loss": 3.2248,
"step": 1380
},
{
"epoch": 4.53,
"learning_rate": 0.0003475,
"loss": 3.26,
"step": 1390
},
{
"epoch": 4.56,
"learning_rate": 0.00035,
"loss": 3.2288,
"step": 1400
},
{
"epoch": 4.56,
"eval_bp": 1.0,
"eval_counts": [
26211,
17123,
12060,
8793
],
"eval_loss": 1.7274117469787598,
"eval_precisions": [
58.21691135641783,
41.200673724735324,
31.649389843852514,
25.342979017754207
],
"eval_ref_len": 41379,
"eval_runtime": 140.7961,
"eval_samples_per_second": 24.596,
"eval_score": 37.242954038765475,
"eval_steps_per_second": 24.596,
"eval_sys_len": 45023,
"eval_totals": [
45023,
41560,
38105,
34696
],
"step": 1400
},
{
"epoch": 4.59,
"learning_rate": 0.0003525,
"loss": 3.2446,
"step": 1410
},
{
"epoch": 4.62,
"learning_rate": 0.000355,
"loss": 3.1755,
"step": 1420
},
{
"epoch": 4.66,
"learning_rate": 0.0003575,
"loss": 3.274,
"step": 1430
},
{
"epoch": 4.69,
"learning_rate": 0.00035999999999999997,
"loss": 3.1518,
"step": 1440
},
{
"epoch": 4.72,
"learning_rate": 0.0003625,
"loss": 3.2472,
"step": 1450
},
{
"epoch": 4.75,
"learning_rate": 0.000365,
"loss": 3.231,
"step": 1460
},
{
"epoch": 4.79,
"learning_rate": 0.0003675,
"loss": 3.2377,
"step": 1470
},
{
"epoch": 4.82,
"learning_rate": 0.00037,
"loss": 3.2576,
"step": 1480
},
{
"epoch": 4.85,
"learning_rate": 0.0003725,
"loss": 3.2433,
"step": 1490
},
{
"epoch": 4.88,
"learning_rate": 0.000375,
"loss": 3.2144,
"step": 1500
},
{
"epoch": 4.92,
"learning_rate": 0.0003775,
"loss": 3.2166,
"step": 1510
},
{
"epoch": 4.95,
"learning_rate": 0.00038,
"loss": 3.1721,
"step": 1520
},
{
"epoch": 4.98,
"learning_rate": 0.00038250000000000003,
"loss": 3.1729,
"step": 1530
},
{
"epoch": 5.01,
"learning_rate": 0.00038500000000000003,
"loss": 3.1772,
"step": 1540
},
{
"epoch": 5.05,
"learning_rate": 0.00038750000000000004,
"loss": 3.1715,
"step": 1550
},
{
"epoch": 5.08,
"learning_rate": 0.00039000000000000005,
"loss": 3.0745,
"step": 1560
},
{
"epoch": 5.11,
"learning_rate": 0.0003925,
"loss": 3.1181,
"step": 1570
},
{
"epoch": 5.14,
"learning_rate": 0.000395,
"loss": 3.0892,
"step": 1580
},
{
"epoch": 5.18,
"learning_rate": 0.0003975,
"loss": 3.1591,
"step": 1590
},
{
"epoch": 5.21,
"learning_rate": 0.0004,
"loss": 3.1232,
"step": 1600
},
{
"epoch": 5.21,
"eval_bp": 1.0,
"eval_counts": [
25985,
17105,
12124,
8931
],
"eval_loss": 1.6549996137619019,
"eval_precisions": [
61.648873072360615,
44.2138185953938,
34.410921579201315,
28.063725490196077
],
"eval_ref_len": 41379,
"eval_runtime": 118.8752,
"eval_samples_per_second": 29.131,
"eval_score": 40.27923361043648,
"eval_steps_per_second": 29.131,
"eval_sys_len": 42150,
"eval_totals": [
42150,
38687,
35233,
31824
],
"step": 1600
},
{
"epoch": 5.24,
"learning_rate": 0.0004025,
"loss": 3.049,
"step": 1610
},
{
"epoch": 5.27,
"learning_rate": 0.00040500000000000003,
"loss": 3.16,
"step": 1620
},
{
"epoch": 5.31,
"learning_rate": 0.0004075,
"loss": 3.1724,
"step": 1630
},
{
"epoch": 5.34,
"learning_rate": 0.00041,
"loss": 3.114,
"step": 1640
},
{
"epoch": 5.37,
"learning_rate": 0.0004125,
"loss": 3.1586,
"step": 1650
},
{
"epoch": 5.4,
"learning_rate": 0.000415,
"loss": 3.0389,
"step": 1660
},
{
"epoch": 5.44,
"learning_rate": 0.0004175,
"loss": 3.0599,
"step": 1670
},
{
"epoch": 5.47,
"learning_rate": 0.00042,
"loss": 3.1169,
"step": 1680
},
{
"epoch": 5.5,
"learning_rate": 0.00042249999999999997,
"loss": 3.0748,
"step": 1690
},
{
"epoch": 5.53,
"learning_rate": 0.000425,
"loss": 3.0661,
"step": 1700
},
{
"epoch": 5.57,
"learning_rate": 0.0004275,
"loss": 3.125,
"step": 1710
},
{
"epoch": 5.6,
"learning_rate": 0.00043,
"loss": 3.153,
"step": 1720
},
{
"epoch": 5.63,
"learning_rate": 0.0004325,
"loss": 3.0269,
"step": 1730
},
{
"epoch": 5.66,
"learning_rate": 0.000435,
"loss": 3.0262,
"step": 1740
},
{
"epoch": 5.7,
"learning_rate": 0.0004375,
"loss": 2.993,
"step": 1750
},
{
"epoch": 5.73,
"learning_rate": 0.00044,
"loss": 3.0263,
"step": 1760
},
{
"epoch": 5.76,
"learning_rate": 0.0004425,
"loss": 2.959,
"step": 1770
},
{
"epoch": 5.79,
"learning_rate": 0.00044500000000000003,
"loss": 3.0096,
"step": 1780
},
{
"epoch": 5.83,
"learning_rate": 0.00044750000000000004,
"loss": 3.0877,
"step": 1790
},
{
"epoch": 5.86,
"learning_rate": 0.00045000000000000004,
"loss": 3.0788,
"step": 1800
},
{
"epoch": 5.86,
"eval_bp": 1.0,
"eval_counts": [
25630,
16905,
12013,
8850
],
"eval_loss": 1.6167092323303223,
"eval_precisions": [
60.873076192285765,
43.748867782924876,
34.141419882907975,
27.850332001132895
],
"eval_ref_len": 41379,
"eval_runtime": 122.6416,
"eval_samples_per_second": 28.237,
"eval_score": 39.89109940073679,
"eval_steps_per_second": 28.237,
"eval_sys_len": 42104,
"eval_totals": [
42104,
38641,
35186,
31777
],
"step": 1800
},
{
"epoch": 5.89,
"learning_rate": 0.00045250000000000005,
"loss": 3.1263,
"step": 1810
},
{
"epoch": 5.93,
"learning_rate": 0.000455,
"loss": 3.0045,
"step": 1820
},
{
"epoch": 5.96,
"learning_rate": 0.0004575,
"loss": 3.0452,
"step": 1830
},
{
"epoch": 5.99,
"learning_rate": 0.00046,
"loss": 3.0684,
"step": 1840
},
{
"epoch": 6.02,
"learning_rate": 0.0004625,
"loss": 3.0385,
"step": 1850
},
{
"epoch": 6.06,
"learning_rate": 0.000465,
"loss": 2.9502,
"step": 1860
},
{
"epoch": 6.09,
"learning_rate": 0.00046750000000000003,
"loss": 3.019,
"step": 1870
},
{
"epoch": 6.12,
"learning_rate": 0.00047,
"loss": 2.9622,
"step": 1880
},
{
"epoch": 6.15,
"learning_rate": 0.0004725,
"loss": 2.9952,
"step": 1890
},
{
"epoch": 6.19,
"learning_rate": 0.000475,
"loss": 2.961,
"step": 1900
},
{
"epoch": 6.22,
"learning_rate": 0.0004775,
"loss": 2.9056,
"step": 1910
},
{
"epoch": 6.25,
"learning_rate": 0.00048,
"loss": 3.03,
"step": 1920
},
{
"epoch": 6.28,
"learning_rate": 0.0004825,
"loss": 2.9475,
"step": 1930
},
{
"epoch": 6.32,
"learning_rate": 0.00048499999999999997,
"loss": 2.9383,
"step": 1940
},
{
"epoch": 6.35,
"learning_rate": 0.0004875,
"loss": 2.9497,
"step": 1950
},
{
"epoch": 6.38,
"learning_rate": 0.00049,
"loss": 2.9825,
"step": 1960
},
{
"epoch": 6.41,
"learning_rate": 0.0004925,
"loss": 2.8683,
"step": 1970
},
{
"epoch": 6.45,
"learning_rate": 0.000495,
"loss": 2.9585,
"step": 1980
},
{
"epoch": 6.48,
"learning_rate": 0.0004975,
"loss": 3.0039,
"step": 1990
},
{
"epoch": 6.51,
"learning_rate": 0.0005,
"loss": 2.8811,
"step": 2000
},
{
"epoch": 6.51,
"eval_bp": 1.0,
"eval_counts": [
27043,
18237,
13141,
9796
],
"eval_loss": 1.6076596975326538,
"eval_precisions": [
62.29957611500184,
45.6552760045062,
36.01359313765792,
29.613954472626137
],
"eval_ref_len": 41379,
"eval_runtime": 127.9727,
"eval_samples_per_second": 27.06,
"eval_score": 41.733492512975715,
"eval_steps_per_second": 27.06,
"eval_sys_len": 43408,
"eval_totals": [
43408,
39945,
36489,
33079
],
"step": 2000
},
{
"epoch": 6.54,
"learning_rate": 0.0005024999999999999,
"loss": 2.9861,
"step": 2010
},
{
"epoch": 6.58,
"learning_rate": 0.000505,
"loss": 3.0033,
"step": 2020
},
{
"epoch": 6.61,
"learning_rate": 0.0005074999999999999,
"loss": 2.8877,
"step": 2030
},
{
"epoch": 6.64,
"learning_rate": 0.00051,
"loss": 2.9709,
"step": 2040
},
{
"epoch": 6.67,
"learning_rate": 0.0005124999999999999,
"loss": 2.8549,
"step": 2050
},
{
"epoch": 6.71,
"learning_rate": 0.000515,
"loss": 2.9543,
"step": 2060
},
{
"epoch": 6.74,
"learning_rate": 0.0005175,
"loss": 2.9383,
"step": 2070
},
{
"epoch": 6.77,
"learning_rate": 0.0005200000000000001,
"loss": 2.8779,
"step": 2080
},
{
"epoch": 6.8,
"learning_rate": 0.0005225,
"loss": 2.969,
"step": 2090
},
{
"epoch": 6.84,
"learning_rate": 0.0005250000000000001,
"loss": 2.968,
"step": 2100
},
{
"epoch": 6.87,
"learning_rate": 0.0005275,
"loss": 2.8835,
"step": 2110
},
{
"epoch": 6.9,
"learning_rate": 0.0005300000000000001,
"loss": 3.0172,
"step": 2120
},
{
"epoch": 6.93,
"learning_rate": 0.0005325,
"loss": 2.9274,
"step": 2130
},
{
"epoch": 6.97,
"learning_rate": 0.000535,
"loss": 2.9879,
"step": 2140
},
{
"epoch": 7.0,
"learning_rate": 0.0005375,
"loss": 2.9297,
"step": 2150
},
{
"epoch": 7.03,
"learning_rate": 0.00054,
"loss": 2.8153,
"step": 2160
},
{
"epoch": 7.06,
"learning_rate": 0.0005425,
"loss": 2.8727,
"step": 2170
},
{
"epoch": 7.1,
"learning_rate": 0.000545,
"loss": 2.8239,
"step": 2180
},
{
"epoch": 7.13,
"learning_rate": 0.0005475,
"loss": 2.9188,
"step": 2190
},
{
"epoch": 7.16,
"learning_rate": 0.00055,
"loss": 2.8291,
"step": 2200
},
{
"epoch": 7.16,
"eval_bp": 1.0,
"eval_counts": [
26097,
17579,
12717,
9479
],
"eval_loss": 1.545904278755188,
"eval_precisions": [
55.22005924672027,
40.13745233691805,
31.5221971593585,
25.666089028484784
],
"eval_ref_len": 41379,
"eval_runtime": 165.394,
"eval_samples_per_second": 20.938,
"eval_score": 36.59364849906923,
"eval_steps_per_second": 20.938,
"eval_sys_len": 47260,
"eval_totals": [
47260,
43797,
40343,
36932
],
"step": 2200
},
{
"epoch": 7.19,
"learning_rate": 0.0005525,
"loss": 2.8052,
"step": 2210
},
{
"epoch": 7.23,
"learning_rate": 0.000555,
"loss": 2.8976,
"step": 2220
},
{
"epoch": 7.26,
"learning_rate": 0.0005575,
"loss": 2.8193,
"step": 2230
},
{
"epoch": 7.29,
"learning_rate": 0.0005600000000000001,
"loss": 2.9084,
"step": 2240
},
{
"epoch": 7.33,
"learning_rate": 0.0005625000000000001,
"loss": 2.8175,
"step": 2250
},
{
"epoch": 7.36,
"learning_rate": 0.000565,
"loss": 2.8522,
"step": 2260
},
{
"epoch": 7.39,
"learning_rate": 0.0005675,
"loss": 2.8453,
"step": 2270
},
{
"epoch": 7.42,
"learning_rate": 0.00057,
"loss": 2.9098,
"step": 2280
},
{
"epoch": 7.46,
"learning_rate": 0.0005725,
"loss": 2.7735,
"step": 2290
},
{
"epoch": 7.49,
"learning_rate": 0.000575,
"loss": 2.8275,
"step": 2300
},
{
"epoch": 7.52,
"learning_rate": 0.0005775,
"loss": 2.8569,
"step": 2310
},
{
"epoch": 7.55,
"learning_rate": 0.00058,
"loss": 2.7792,
"step": 2320
},
{
"epoch": 7.59,
"learning_rate": 0.0005825,
"loss": 2.8173,
"step": 2330
},
{
"epoch": 7.62,
"learning_rate": 0.000585,
"loss": 2.8759,
"step": 2340
},
{
"epoch": 7.65,
"learning_rate": 0.0005875,
"loss": 2.8538,
"step": 2350
},
{
"epoch": 7.68,
"learning_rate": 0.00059,
"loss": 2.902,
"step": 2360
},
{
"epoch": 7.72,
"learning_rate": 0.0005925,
"loss": 2.7693,
"step": 2370
},
{
"epoch": 7.75,
"learning_rate": 0.0005949999999999999,
"loss": 2.8489,
"step": 2380
},
{
"epoch": 7.78,
"learning_rate": 0.0005975,
"loss": 2.8515,
"step": 2390
},
{
"epoch": 7.81,
"learning_rate": 0.0006,
"loss": 2.8358,
"step": 2400
},
{
"epoch": 7.81,
"eval_bp": 0.951493698303458,
"eval_counts": [
27661,
18999,
13858,
10389
],
"eval_loss": 1.5499262809753418,
"eval_precisions": [
70.17174459017225,
52.839581710980084,
42.60722521137586,
35.65324822402965
],
"eval_ref_len": 41379,
"eval_runtime": 72.495,
"eval_samples_per_second": 47.769,
"eval_score": 46.35343428916555,
"eval_steps_per_second": 47.769,
"eval_sys_len": 39419,
"eval_totals": [
39419,
35956,
32525,
29139
],
"step": 2400
},
{
"epoch": 7.85,
"learning_rate": 0.0006025000000000001,
"loss": 2.8804,
"step": 2410
},
{
"epoch": 7.88,
"learning_rate": 0.000605,
"loss": 2.7958,
"step": 2420
},
{
"epoch": 7.91,
"learning_rate": 0.0006075000000000001,
"loss": 2.8646,
"step": 2430
},
{
"epoch": 7.94,
"learning_rate": 0.00061,
"loss": 2.7276,
"step": 2440
},
{
"epoch": 7.98,
"learning_rate": 0.0006125000000000001,
"loss": 2.8928,
"step": 2450
},
{
"epoch": 8.01,
"learning_rate": 0.000615,
"loss": 2.8545,
"step": 2460
},
{
"epoch": 8.04,
"learning_rate": 0.0006175000000000001,
"loss": 2.7296,
"step": 2470
},
{
"epoch": 8.07,
"learning_rate": 0.00062,
"loss": 2.7535,
"step": 2480
},
{
"epoch": 8.11,
"learning_rate": 0.0006225000000000001,
"loss": 2.8058,
"step": 2490
},
{
"epoch": 8.14,
"learning_rate": 0.000625,
"loss": 2.7666,
"step": 2500
},
{
"epoch": 8.17,
"learning_rate": 0.0006274999999999999,
"loss": 2.7758,
"step": 2510
},
{
"epoch": 8.2,
"learning_rate": 0.00063,
"loss": 2.6975,
"step": 2520
},
{
"epoch": 8.24,
"learning_rate": 0.0006324999999999999,
"loss": 2.7532,
"step": 2530
},
{
"epoch": 8.27,
"learning_rate": 0.000635,
"loss": 2.7222,
"step": 2540
},
{
"epoch": 8.3,
"learning_rate": 0.0006374999999999999,
"loss": 2.8252,
"step": 2550
},
{
"epoch": 8.33,
"learning_rate": 0.00064,
"loss": 2.8289,
"step": 2560
},
{
"epoch": 8.37,
"learning_rate": 0.0006425,
"loss": 2.7046,
"step": 2570
},
{
"epoch": 8.4,
"learning_rate": 0.0006450000000000001,
"loss": 2.7199,
"step": 2580
},
{
"epoch": 8.43,
"learning_rate": 0.0006475,
"loss": 2.6622,
"step": 2590
},
{
"epoch": 8.46,
"learning_rate": 0.0006500000000000001,
"loss": 2.764,
"step": 2600
},
{
"epoch": 8.46,
"eval_bp": 1.0,
"eval_counts": [
26024,
17669,
12817,
9561
],
"eval_loss": 1.508206844329834,
"eval_precisions": [
55.35019248356977,
40.56803049088488,
31.96418774003691,
26.063133791298657
],
"eval_ref_len": 41379,
"eval_runtime": 173.7893,
"eval_samples_per_second": 19.926,
"eval_score": 36.9826767402354,
"eval_steps_per_second": 19.926,
"eval_sys_len": 47017,
"eval_totals": [
47017,
43554,
40098,
36684
],
"step": 2600
},
{
"epoch": 8.5,
"learning_rate": 0.0006525,
"loss": 2.7553,
"step": 2610
},
{
"epoch": 8.53,
"learning_rate": 0.0006550000000000001,
"loss": 2.7883,
"step": 2620
},
{
"epoch": 8.56,
"learning_rate": 0.0006575,
"loss": 2.7572,
"step": 2630
},
{
"epoch": 8.59,
"learning_rate": 0.00066,
"loss": 2.7026,
"step": 2640
},
{
"epoch": 8.63,
"learning_rate": 0.0006625,
"loss": 2.7265,
"step": 2650
},
{
"epoch": 8.66,
"learning_rate": 0.000665,
"loss": 2.5958,
"step": 2660
},
{
"epoch": 8.69,
"learning_rate": 0.0006675,
"loss": 2.69,
"step": 2670
},
{
"epoch": 8.72,
"learning_rate": 0.00067,
"loss": 2.7464,
"step": 2680
},
{
"epoch": 8.76,
"learning_rate": 0.0006725,
"loss": 2.7263,
"step": 2690
},
{
"epoch": 8.79,
"learning_rate": 0.000675,
"loss": 2.7336,
"step": 2700
},
{
"epoch": 8.82,
"learning_rate": 0.0006775,
"loss": 2.7562,
"step": 2710
},
{
"epoch": 8.86,
"learning_rate": 0.00068,
"loss": 2.7315,
"step": 2720
},
{
"epoch": 8.89,
"learning_rate": 0.0006825000000000001,
"loss": 2.7706,
"step": 2730
},
{
"epoch": 8.92,
"learning_rate": 0.0006850000000000001,
"loss": 2.6668,
"step": 2740
},
{
"epoch": 8.95,
"learning_rate": 0.0006875,
"loss": 2.7939,
"step": 2750
},
{
"epoch": 8.99,
"learning_rate": 0.00069,
"loss": 2.7894,
"step": 2760
},
{
"epoch": 9.02,
"learning_rate": 0.0006925,
"loss": 2.7196,
"step": 2770
},
{
"epoch": 9.05,
"learning_rate": 0.000695,
"loss": 2.6768,
"step": 2780
},
{
"epoch": 9.08,
"learning_rate": 0.0006975,
"loss": 2.5136,
"step": 2790
},
{
"epoch": 9.12,
"learning_rate": 0.0007,
"loss": 2.6378,
"step": 2800
},
{
"epoch": 9.12,
"eval_bp": 0.982521383577959,
"eval_counts": [
27410,
18929,
13880,
10452
],
"eval_loss": 1.4984954595565796,
"eval_precisions": [
67.40937484629384,
50.88577649936826,
41.13445751711466,
34.460929772502475
],
"eval_ref_len": 41379,
"eval_runtime": 88.8496,
"eval_samples_per_second": 38.976,
"eval_score": 46.13753405463289,
"eval_steps_per_second": 38.976,
"eval_sys_len": 40662,
"eval_totals": [
40662,
37199,
33743,
30330
],
"step": 2800
},
{
"epoch": 9.15,
"learning_rate": 0.0007025,
"loss": 2.6048,
"step": 2810
},
{
"epoch": 9.18,
"learning_rate": 0.000705,
"loss": 2.6136,
"step": 2820
},
{
"epoch": 9.21,
"learning_rate": 0.0007075,
"loss": 2.5976,
"step": 2830
},
{
"epoch": 9.25,
"learning_rate": 0.00071,
"loss": 2.6645,
"step": 2840
},
{
"epoch": 9.28,
"learning_rate": 0.0007125,
"loss": 2.7466,
"step": 2850
},
{
"epoch": 9.31,
"learning_rate": 0.000715,
"loss": 2.6573,
"step": 2860
},
{
"epoch": 9.34,
"learning_rate": 0.0007175,
"loss": 2.6746,
"step": 2870
},
{
"epoch": 9.38,
"learning_rate": 0.0007199999999999999,
"loss": 2.6083,
"step": 2880
},
{
"epoch": 9.41,
"learning_rate": 0.0007225,
"loss": 2.718,
"step": 2890
},
{
"epoch": 9.44,
"learning_rate": 0.000725,
"loss": 2.6234,
"step": 2900
},
{
"epoch": 9.47,
"learning_rate": 0.0007275000000000001,
"loss": 2.6455,
"step": 2910
},
{
"epoch": 9.51,
"learning_rate": 0.00073,
"loss": 2.6674,
"step": 2920
},
{
"epoch": 9.54,
"learning_rate": 0.0007325000000000001,
"loss": 2.6671,
"step": 2930
},
{
"epoch": 9.57,
"learning_rate": 0.000735,
"loss": 2.6839,
"step": 2940
},
{
"epoch": 9.6,
"learning_rate": 0.0007375000000000001,
"loss": 2.7052,
"step": 2950
},
{
"epoch": 9.64,
"learning_rate": 0.00074,
"loss": 2.7401,
"step": 2960
},
{
"epoch": 9.67,
"learning_rate": 0.0007425000000000001,
"loss": 2.6638,
"step": 2970
},
{
"epoch": 9.7,
"learning_rate": 0.000745,
"loss": 2.5944,
"step": 2980
},
{
"epoch": 9.73,
"learning_rate": 0.0007475000000000001,
"loss": 2.6262,
"step": 2990
},
{
"epoch": 9.77,
"learning_rate": 0.00075,
"loss": 2.701,
"step": 3000
},
{
"epoch": 9.77,
"eval_bp": 1.0,
"eval_counts": [
26685,
18421,
13517,
10181
],
"eval_loss": 1.4741408824920654,
"eval_precisions": [
63.57355568790947,
47.83184461985874,
38.558306709265175,
32.17457257529311
],
"eval_ref_len": 41379,
"eval_runtime": 129.3598,
"eval_samples_per_second": 26.77,
"eval_score": 44.071323498930454,
"eval_steps_per_second": 26.77,
"eval_sys_len": 41975,
"eval_totals": [
41975,
38512,
35056,
31643
],
"step": 3000
},
{
"epoch": 9.8,
"learning_rate": 0.0007524999999999999,
"loss": 2.6454,
"step": 3010
},
{
"epoch": 9.83,
"learning_rate": 0.000755,
"loss": 2.7318,
"step": 3020
},
{
"epoch": 9.86,
"learning_rate": 0.0007574999999999999,
"loss": 2.6678,
"step": 3030
},
{
"epoch": 9.9,
"learning_rate": 0.00076,
"loss": 2.6132,
"step": 3040
},
{
"epoch": 9.93,
"learning_rate": 0.0007624999999999999,
"loss": 2.6956,
"step": 3050
},
{
"epoch": 9.96,
"learning_rate": 0.0007650000000000001,
"loss": 2.5574,
"step": 3060
},
{
"epoch": 9.99,
"learning_rate": 0.0007675,
"loss": 2.6114,
"step": 3070
},
{
"epoch": 10.03,
"learning_rate": 0.0007700000000000001,
"loss": 2.6667,
"step": 3080
},
{
"epoch": 10.06,
"learning_rate": 0.0007725,
"loss": 2.5687,
"step": 3090
},
{
"epoch": 10.09,
"learning_rate": 0.0007750000000000001,
"loss": 2.4745,
"step": 3100
},
{
"epoch": 10.12,
"learning_rate": 0.0007775,
"loss": 2.6286,
"step": 3110
},
{
"epoch": 10.16,
"learning_rate": 0.0007800000000000001,
"loss": 2.6025,
"step": 3120
},
{
"epoch": 10.19,
"learning_rate": 0.0007825,
"loss": 2.5623,
"step": 3130
},
{
"epoch": 10.22,
"learning_rate": 0.000785,
"loss": 2.6493,
"step": 3140
},
{
"epoch": 10.26,
"learning_rate": 0.0007875,
"loss": 2.5866,
"step": 3150
},
{
"epoch": 10.29,
"learning_rate": 0.00079,
"loss": 2.5541,
"step": 3160
},
{
"epoch": 10.32,
"learning_rate": 0.0007925,
"loss": 2.6685,
"step": 3170
},
{
"epoch": 10.35,
"learning_rate": 0.000795,
"loss": 2.5534,
"step": 3180
},
{
"epoch": 10.39,
"learning_rate": 0.0007975,
"loss": 2.6261,
"step": 3190
},
{
"epoch": 10.42,
"learning_rate": 0.0008,
"loss": 2.6547,
"step": 3200
},
{
"epoch": 10.42,
"eval_bp": 0.9986941383168589,
"eval_counts": [
27057,
18731,
13763,
10375
],
"eval_loss": 1.4648829698562622,
"eval_precisions": [
65.47368421052632,
49.47176588664096,
40.00174388188107,
33.47422081693231
],
"eval_ref_len": 41379,
"eval_runtime": 105.6078,
"eval_samples_per_second": 32.791,
"eval_score": 45.57596776445642,
"eval_steps_per_second": 32.791,
"eval_sys_len": 41325,
"eval_totals": [
41325,
37862,
34406,
30994
],
"step": 3200
},
{
"epoch": 10.45,
"learning_rate": 0.0008025,
"loss": 2.4607,
"step": 3210
},
{
"epoch": 10.48,
"learning_rate": 0.000805,
"loss": 2.596,
"step": 3220
},
{
"epoch": 10.52,
"learning_rate": 0.0008075000000000001,
"loss": 2.5666,
"step": 3230
},
{
"epoch": 10.55,
"learning_rate": 0.0008100000000000001,
"loss": 2.5642,
"step": 3240
},
{
"epoch": 10.58,
"learning_rate": 0.0008125000000000001,
"loss": 2.5876,
"step": 3250
},
{
"epoch": 10.61,
"learning_rate": 0.000815,
"loss": 2.5194,
"step": 3260
},
{
"epoch": 10.65,
"learning_rate": 0.0008175,
"loss": 2.71,
"step": 3270
},
{
"epoch": 10.68,
"learning_rate": 0.00082,
"loss": 2.5289,
"step": 3280
},
{
"epoch": 10.71,
"learning_rate": 0.0008225,
"loss": 2.5532,
"step": 3290
},
{
"epoch": 10.74,
"learning_rate": 0.000825,
"loss": 2.5693,
"step": 3300
},
{
"epoch": 10.78,
"learning_rate": 0.0008275,
"loss": 2.5581,
"step": 3310
},
{
"epoch": 10.81,
"learning_rate": 0.00083,
"loss": 2.5051,
"step": 3320
},
{
"epoch": 10.84,
"learning_rate": 0.0008325,
"loss": 2.5048,
"step": 3330
},
{
"epoch": 10.87,
"learning_rate": 0.000835,
"loss": 2.5013,
"step": 3340
},
{
"epoch": 10.91,
"learning_rate": 0.0008375,
"loss": 2.5961,
"step": 3350
},
{
"epoch": 10.94,
"learning_rate": 0.00084,
"loss": 2.5747,
"step": 3360
},
{
"epoch": 10.97,
"learning_rate": 0.0008425,
"loss": 2.513,
"step": 3370
},
{
"epoch": 11.0,
"learning_rate": 0.0008449999999999999,
"loss": 2.5362,
"step": 3380
},
{
"epoch": 11.04,
"learning_rate": 0.0008475000000000001,
"loss": 2.4747,
"step": 3390
},
{
"epoch": 11.07,
"learning_rate": 0.00085,
"loss": 2.4868,
"step": 3400
},
{
"epoch": 11.07,
"eval_bp": 1.0,
"eval_counts": [
27887,
19470,
14351,
10835
],
"eval_loss": 1.463901162147522,
"eval_precisions": [
65.5994918962151,
49.861708666256916,
40.29255692506388,
33.617747440273035
],
"eval_ref_len": 41379,
"eval_runtime": 111.3022,
"eval_samples_per_second": 31.113,
"eval_score": 45.87913991332588,
"eval_steps_per_second": 31.113,
"eval_sys_len": 42511,
"eval_totals": [
42511,
39048,
35617,
32230
],
"step": 3400
},
{
"epoch": 11.1,
"learning_rate": 0.0008525000000000001,
"loss": 2.4984,
"step": 3410
},
{
"epoch": 11.13,
"learning_rate": 0.000855,
"loss": 2.4708,
"step": 3420
},
{
"epoch": 11.17,
"learning_rate": 0.0008575000000000001,
"loss": 2.4994,
"step": 3430
},
{
"epoch": 11.2,
"learning_rate": 0.00086,
"loss": 2.3728,
"step": 3440
},
{
"epoch": 11.23,
"learning_rate": 0.0008625000000000001,
"loss": 2.4748,
"step": 3450
},
{
"epoch": 11.26,
"learning_rate": 0.000865,
"loss": 2.4959,
"step": 3460
},
{
"epoch": 11.3,
"learning_rate": 0.0008675000000000001,
"loss": 2.513,
"step": 3470
},
{
"epoch": 11.33,
"learning_rate": 0.00087,
"loss": 2.5693,
"step": 3480
},
{
"epoch": 11.36,
"learning_rate": 0.0008725000000000001,
"loss": 2.5396,
"step": 3490
},
{
"epoch": 11.39,
"learning_rate": 0.000875,
"loss": 2.5051,
"step": 3500
},
{
"epoch": 11.43,
"learning_rate": 0.0008774999999999999,
"loss": 2.4765,
"step": 3510
},
{
"epoch": 11.46,
"learning_rate": 0.00088,
"loss": 2.4833,
"step": 3520
},
{
"epoch": 11.49,
"learning_rate": 0.0008824999999999999,
"loss": 2.509,
"step": 3530
},
{
"epoch": 11.52,
"learning_rate": 0.000885,
"loss": 2.5082,
"step": 3540
},
{
"epoch": 11.56,
"learning_rate": 0.0008874999999999999,
"loss": 2.5851,
"step": 3550
},
{
"epoch": 11.59,
"learning_rate": 0.0008900000000000001,
"loss": 2.4521,
"step": 3560
},
{
"epoch": 11.62,
"learning_rate": 0.0008925,
"loss": 2.4873,
"step": 3570
},
{
"epoch": 11.65,
"learning_rate": 0.0008950000000000001,
"loss": 2.4671,
"step": 3580
},
{
"epoch": 11.69,
"learning_rate": 0.0008975,
"loss": 2.5156,
"step": 3590
},
{
"epoch": 11.72,
"learning_rate": 0.0009000000000000001,
"loss": 2.4155,
"step": 3600
},
{
"epoch": 11.72,
"eval_bp": 1.0,
"eval_counts": [
26675,
18446,
13573,
10248
],
"eval_loss": 1.4466536045074463,
"eval_precisions": [
64.46350894151765,
48.64836353087006,
39.38655291488929,
33.009083295754685
],
"eval_ref_len": 41379,
"eval_runtime": 110.662,
"eval_samples_per_second": 31.293,
"eval_score": 44.93563710796045,
"eval_steps_per_second": 31.293,
"eval_sys_len": 41380,
"eval_totals": [
41380,
37917,
34461,
31046
],
"step": 3600
},
{
"epoch": 11.75,
"learning_rate": 0.0009025,
"loss": 2.5093,
"step": 3610
},
{
"epoch": 11.79,
"learning_rate": 0.0009050000000000001,
"loss": 2.532,
"step": 3620
},
{
"epoch": 11.82,
"learning_rate": 0.0009075,
"loss": 2.5547,
"step": 3630
},
{
"epoch": 11.85,
"learning_rate": 0.00091,
"loss": 2.3751,
"step": 3640
},
{
"epoch": 11.88,
"learning_rate": 0.0009125,
"loss": 2.5222,
"step": 3650
},
{
"epoch": 11.92,
"learning_rate": 0.000915,
"loss": 2.4935,
"step": 3660
},
{
"epoch": 11.95,
"learning_rate": 0.0009175,
"loss": 2.5756,
"step": 3670
},
{
"epoch": 11.98,
"learning_rate": 0.00092,
"loss": 2.5278,
"step": 3680
},
{
"epoch": 12.01,
"learning_rate": 0.0009225,
"loss": 2.4911,
"step": 3690
},
{
"epoch": 12.05,
"learning_rate": 0.000925,
"loss": 2.4734,
"step": 3700
},
{
"epoch": 12.08,
"learning_rate": 0.0009275,
"loss": 2.3916,
"step": 3710
},
{
"epoch": 12.11,
"learning_rate": 0.00093,
"loss": 2.443,
"step": 3720
},
{
"epoch": 12.14,
"learning_rate": 0.0009325000000000001,
"loss": 2.3527,
"step": 3730
},
{
"epoch": 12.18,
"learning_rate": 0.0009350000000000001,
"loss": 2.3896,
"step": 3740
},
{
"epoch": 12.21,
"learning_rate": 0.0009375,
"loss": 2.4948,
"step": 3750
},
{
"epoch": 12.24,
"learning_rate": 0.00094,
"loss": 2.4026,
"step": 3760
},
{
"epoch": 12.27,
"learning_rate": 0.0009425,
"loss": 2.3684,
"step": 3770
},
{
"epoch": 12.31,
"learning_rate": 0.000945,
"loss": 2.434,
"step": 3780
},
{
"epoch": 12.34,
"learning_rate": 0.0009475,
"loss": 2.426,
"step": 3790
},
{
"epoch": 12.37,
"learning_rate": 0.00095,
"loss": 2.4975,
"step": 3800
},
{
"epoch": 12.37,
"eval_bp": 1.0,
"eval_counts": [
26470,
18231,
13333,
10001
],
"eval_loss": 1.4460936784744263,
"eval_precisions": [
56.14593276063209,
41.73572638615448,
33.11888320333847,
27.118414273705902
],
"eval_ref_len": 41379,
"eval_runtime": 158.852,
"eval_samples_per_second": 21.8,
"eval_score": 38.08829771113567,
"eval_steps_per_second": 21.8,
"eval_sys_len": 47145,
"eval_totals": [
47145,
43682,
40258,
36879
],
"step": 3800
},
{
"epoch": 12.4,
"learning_rate": 0.0009525,
"loss": 2.4968,
"step": 3810
},
{
"epoch": 12.44,
"learning_rate": 0.000955,
"loss": 2.4392,
"step": 3820
},
{
"epoch": 12.47,
"learning_rate": 0.0009575,
"loss": 2.4044,
"step": 3830
},
{
"epoch": 12.5,
"learning_rate": 0.00096,
"loss": 2.4352,
"step": 3840
},
{
"epoch": 12.53,
"learning_rate": 0.0009625,
"loss": 2.398,
"step": 3850
},
{
"epoch": 12.57,
"learning_rate": 0.000965,
"loss": 2.5434,
"step": 3860
},
{
"epoch": 12.6,
"learning_rate": 0.0009675,
"loss": 2.335,
"step": 3870
},
{
"epoch": 12.63,
"learning_rate": 0.0009699999999999999,
"loss": 2.41,
"step": 3880
},
{
"epoch": 12.66,
"learning_rate": 0.0009725000000000001,
"loss": 2.4202,
"step": 3890
},
{
"epoch": 12.7,
"learning_rate": 0.000975,
"loss": 2.4073,
"step": 3900
},
{
"epoch": 12.73,
"learning_rate": 0.0009775,
"loss": 2.3379,
"step": 3910
},
{
"epoch": 12.76,
"learning_rate": 0.00098,
"loss": 2.4531,
"step": 3920
},
{
"epoch": 12.79,
"learning_rate": 0.0009825,
"loss": 2.3775,
"step": 3930
},
{
"epoch": 12.83,
"learning_rate": 0.000985,
"loss": 2.4836,
"step": 3940
},
{
"epoch": 12.86,
"learning_rate": 0.0009875,
"loss": 2.5058,
"step": 3950
},
{
"epoch": 12.89,
"learning_rate": 0.00099,
"loss": 2.4745,
"step": 3960
},
{
"epoch": 12.92,
"learning_rate": 0.0009925000000000001,
"loss": 2.426,
"step": 3970
},
{
"epoch": 12.96,
"learning_rate": 0.000995,
"loss": 2.4218,
"step": 3980
},
{
"epoch": 12.99,
"learning_rate": 0.0009975000000000001,
"loss": 2.3397,
"step": 3990
},
{
"epoch": 13.02,
"learning_rate": 0.001,
"loss": 2.3011,
"step": 4000
},
{
"epoch": 13.02,
"eval_bp": 0.9755140388093076,
"eval_counts": [
28338,
19935,
14788,
11220
],
"eval_loss": 1.4539415836334229,
"eval_precisions": [
70.18178215860122,
54.00243803331979,
44.164377015888185,
37.280701754385966
],
"eval_ref_len": 41379,
"eval_runtime": 75.0864,
"eval_samples_per_second": 46.12,
"eval_score": 48.756442564740674,
"eval_steps_per_second": 46.12,
"eval_sys_len": 40378,
"eval_totals": [
40378,
36915,
33484,
30096
],
"step": 4000
},
{
"epoch": 13.05,
"learning_rate": 0.001,
"loss": 2.4352,
"step": 4010
},
{
"epoch": 13.09,
"learning_rate": 0.001,
"loss": 2.3339,
"step": 4020
},
{
"epoch": 13.12,
"learning_rate": 0.001,
"loss": 2.3264,
"step": 4030
},
{
"epoch": 13.15,
"learning_rate": 0.001,
"loss": 2.3393,
"step": 4040
},
{
"epoch": 13.19,
"learning_rate": 0.001,
"loss": 2.4056,
"step": 4050
},
{
"epoch": 13.22,
"learning_rate": 0.001,
"loss": 2.3324,
"step": 4060
},
{
"epoch": 13.25,
"learning_rate": 0.001,
"loss": 2.3016,
"step": 4070
},
{
"epoch": 13.28,
"learning_rate": 0.001,
"loss": 2.3232,
"step": 4080
},
{
"epoch": 13.32,
"learning_rate": 0.001,
"loss": 2.3512,
"step": 4090
},
{
"epoch": 13.35,
"learning_rate": 0.001,
"loss": 2.2561,
"step": 4100
},
{
"epoch": 13.38,
"learning_rate": 0.001,
"loss": 2.3663,
"step": 4110
},
{
"epoch": 13.41,
"learning_rate": 0.001,
"loss": 2.414,
"step": 4120
},
{
"epoch": 13.45,
"learning_rate": 0.001,
"loss": 2.3334,
"step": 4130
},
{
"epoch": 13.48,
"learning_rate": 0.001,
"loss": 2.324,
"step": 4140
},
{
"epoch": 13.51,
"learning_rate": 0.001,
"loss": 2.4558,
"step": 4150
},
{
"epoch": 13.54,
"learning_rate": 0.001,
"loss": 2.2729,
"step": 4160
},
{
"epoch": 13.58,
"learning_rate": 0.001,
"loss": 2.3651,
"step": 4170
},
{
"epoch": 13.61,
"learning_rate": 0.001,
"loss": 2.4071,
"step": 4180
},
{
"epoch": 13.64,
"learning_rate": 0.001,
"loss": 2.3455,
"step": 4190
},
{
"epoch": 13.67,
"learning_rate": 0.001,
"loss": 2.383,
"step": 4200
},
{
"epoch": 13.67,
"eval_bp": 1.0,
"eval_counts": [
26856,
18659,
13769,
10407
],
"eval_loss": 1.422828197479248,
"eval_precisions": [
63.941334730125476,
48.41714671233588,
39.242454470316645,
32.85037878787879
],
"eval_ref_len": 41379,
"eval_runtime": 112.127,
"eval_samples_per_second": 30.885,
"eval_score": 44.6960720730811,
"eval_steps_per_second": 30.885,
"eval_sys_len": 42001,
"eval_totals": [
42001,
38538,
35087,
31680
],
"step": 4200
},
{
"epoch": 13.71,
"learning_rate": 0.001,
"loss": 2.2856,
"step": 4210
},
{
"epoch": 13.74,
"learning_rate": 0.001,
"loss": 2.33,
"step": 4220
},
{
"epoch": 13.77,
"learning_rate": 0.001,
"loss": 2.3685,
"step": 4230
},
{
"epoch": 13.8,
"learning_rate": 0.001,
"loss": 2.3132,
"step": 4240
},
{
"epoch": 13.84,
"learning_rate": 0.001,
"loss": 2.3598,
"step": 4250
},
{
"epoch": 13.87,
"learning_rate": 0.001,
"loss": 2.3929,
"step": 4260
},
{
"epoch": 13.9,
"learning_rate": 0.001,
"loss": 2.2815,
"step": 4270
},
{
"epoch": 13.93,
"learning_rate": 0.001,
"loss": 2.3766,
"step": 4280
},
{
"epoch": 13.97,
"learning_rate": 0.001,
"loss": 2.4054,
"step": 4290
},
{
"epoch": 14.0,
"learning_rate": 0.001,
"loss": 2.3955,
"step": 4300
},
{
"epoch": 14.03,
"learning_rate": 0.001,
"loss": 2.3044,
"step": 4310
},
{
"epoch": 14.06,
"learning_rate": 0.001,
"loss": 2.2661,
"step": 4320
},
{
"epoch": 14.1,
"learning_rate": 0.001,
"loss": 2.2011,
"step": 4330
},
{
"epoch": 14.13,
"learning_rate": 0.001,
"loss": 2.2519,
"step": 4340
},
{
"epoch": 14.16,
"learning_rate": 0.001,
"loss": 2.2913,
"step": 4350
},
{
"epoch": 14.19,
"learning_rate": 0.001,
"loss": 2.284,
"step": 4360
},
{
"epoch": 14.23,
"learning_rate": 0.001,
"loss": 2.2612,
"step": 4370
},
{
"epoch": 14.26,
"learning_rate": 0.001,
"loss": 2.2675,
"step": 4380
},
{
"epoch": 14.29,
"learning_rate": 0.001,
"loss": 2.2262,
"step": 4390
},
{
"epoch": 14.32,
"learning_rate": 0.001,
"loss": 2.2696,
"step": 4400
},
{
"epoch": 14.32,
"eval_bp": 0.9583631202356716,
"eval_counts": [
27315,
19078,
14092,
10677
],
"eval_loss": 1.4258805513381958,
"eval_precisions": [
68.81912776196116,
52.660925251186924,
42.99880999603332,
36.36208834247182
],
"eval_ref_len": 41379,
"eval_runtime": 79.9145,
"eval_samples_per_second": 43.334,
"eval_score": 46.757988379346315,
"eval_steps_per_second": 43.334,
"eval_sys_len": 39691,
"eval_totals": [
39691,
36228,
32773,
29363
],
"step": 4400
},
{
"epoch": 14.36,
"learning_rate": 0.001,
"loss": 2.3408,
"step": 4410
},
{
"epoch": 14.39,
"learning_rate": 0.001,
"loss": 2.32,
"step": 4420
},
{
"epoch": 14.42,
"learning_rate": 0.001,
"loss": 2.2473,
"step": 4430
},
{
"epoch": 14.45,
"learning_rate": 0.001,
"loss": 2.243,
"step": 4440
},
{
"epoch": 14.49,
"learning_rate": 0.001,
"loss": 2.294,
"step": 4450
},
{
"epoch": 14.52,
"learning_rate": 0.001,
"loss": 2.2595,
"step": 4460
},
{
"epoch": 14.55,
"learning_rate": 0.001,
"loss": 2.2146,
"step": 4470
},
{
"epoch": 14.58,
"learning_rate": 0.001,
"loss": 2.2468,
"step": 4480
},
{
"epoch": 14.62,
"learning_rate": 0.001,
"loss": 2.3313,
"step": 4490
},
{
"epoch": 14.65,
"learning_rate": 0.001,
"loss": 2.2623,
"step": 4500
},
{
"epoch": 14.68,
"learning_rate": 0.001,
"loss": 2.2167,
"step": 4510
},
{
"epoch": 14.72,
"learning_rate": 0.001,
"loss": 2.3084,
"step": 4520
},
{
"epoch": 14.75,
"learning_rate": 0.001,
"loss": 2.2722,
"step": 4530
},
{
"epoch": 14.78,
"learning_rate": 0.001,
"loss": 2.2878,
"step": 4540
},
{
"epoch": 14.81,
"learning_rate": 0.001,
"loss": 2.2647,
"step": 4550
},
{
"epoch": 14.85,
"learning_rate": 0.001,
"loss": 2.35,
"step": 4560
},
{
"epoch": 14.88,
"learning_rate": 0.001,
"loss": 2.312,
"step": 4570
},
{
"epoch": 14.91,
"learning_rate": 0.001,
"loss": 2.3462,
"step": 4580
},
{
"epoch": 14.94,
"learning_rate": 0.001,
"loss": 2.2526,
"step": 4590
},
{
"epoch": 14.98,
"learning_rate": 0.001,
"loss": 2.2648,
"step": 4600
},
{
"epoch": 14.98,
"eval_bp": 0.96840860582464,
"eval_counts": [
27875,
19603,
14568,
11071
],
"eval_loss": 1.4148167371749878,
"eval_precisions": [
69.52758655093285,
53.5177045510388,
43.91258477769404,
37.19094329481322
],
"eval_ref_len": 41379,
"eval_runtime": 80.5243,
"eval_samples_per_second": 43.006,
"eval_score": 48.08159289033059,
"eval_steps_per_second": 43.006,
"eval_sys_len": 40092,
"eval_totals": [
40092,
36629,
33175,
29768
],
"step": 4600
},
{
"epoch": 15.01,
"learning_rate": 0.001,
"loss": 2.2865,
"step": 4610
},
{
"epoch": 15.04,
"learning_rate": 0.001,
"loss": 2.1537,
"step": 4620
},
{
"epoch": 15.07,
"learning_rate": 0.001,
"loss": 2.1461,
"step": 4630
},
{
"epoch": 15.11,
"learning_rate": 0.001,
"loss": 2.2309,
"step": 4640
},
{
"epoch": 15.14,
"learning_rate": 0.001,
"loss": 2.1519,
"step": 4650
},
{
"epoch": 15.17,
"learning_rate": 0.001,
"loss": 2.2306,
"step": 4660
},
{
"epoch": 15.2,
"learning_rate": 0.001,
"loss": 2.1776,
"step": 4670
},
{
"epoch": 15.24,
"learning_rate": 0.001,
"loss": 2.1811,
"step": 4680
},
{
"epoch": 15.27,
"learning_rate": 0.001,
"loss": 2.2001,
"step": 4690
},
{
"epoch": 15.3,
"learning_rate": 0.001,
"loss": 2.1924,
"step": 4700
},
{
"epoch": 15.33,
"learning_rate": 0.001,
"loss": 2.1418,
"step": 4710
},
{
"epoch": 15.37,
"learning_rate": 0.001,
"loss": 2.2936,
"step": 4720
},
{
"epoch": 15.4,
"learning_rate": 0.001,
"loss": 2.2203,
"step": 4730
},
{
"epoch": 15.43,
"learning_rate": 0.001,
"loss": 2.2136,
"step": 4740
},
{
"epoch": 15.46,
"learning_rate": 0.001,
"loss": 2.1555,
"step": 4750
},
{
"epoch": 15.5,
"learning_rate": 0.001,
"loss": 2.3135,
"step": 4760
},
{
"epoch": 15.53,
"learning_rate": 0.001,
"loss": 2.2464,
"step": 4770
},
{
"epoch": 15.56,
"learning_rate": 0.001,
"loss": 2.2005,
"step": 4780
},
{
"epoch": 15.59,
"learning_rate": 0.001,
"loss": 2.1989,
"step": 4790
},
{
"epoch": 15.63,
"learning_rate": 0.001,
"loss": 2.2123,
"step": 4800
},
{
"epoch": 15.63,
"eval_bp": 1.0,
"eval_counts": [
28009,
19746,
14685,
11165
],
"eval_loss": 1.432946801185608,
"eval_precisions": [
67.3374203630244,
51.78327913563411,
42.3491752220556,
35.71314333237373
],
"eval_ref_len": 41379,
"eval_runtime": 95.3413,
"eval_samples_per_second": 36.322,
"eval_score": 47.92142944979905,
"eval_steps_per_second": 36.322,
"eval_sys_len": 41595,
"eval_totals": [
41595,
38132,
34676,
31263
],
"step": 4800
},
{
"epoch": 15.66,
"learning_rate": 0.001,
"loss": 2.2782,
"step": 4810
},
{
"epoch": 15.69,
"learning_rate": 0.001,
"loss": 2.2205,
"step": 4820
},
{
"epoch": 15.72,
"learning_rate": 0.001,
"loss": 2.212,
"step": 4830
},
{
"epoch": 15.76,
"learning_rate": 0.001,
"loss": 2.2153,
"step": 4840
},
{
"epoch": 15.79,
"learning_rate": 0.001,
"loss": 2.2444,
"step": 4850
},
{
"epoch": 15.82,
"learning_rate": 0.001,
"loss": 2.285,
"step": 4860
},
{
"epoch": 15.85,
"learning_rate": 0.001,
"loss": 2.2001,
"step": 4870
},
{
"epoch": 15.89,
"learning_rate": 0.001,
"loss": 2.3122,
"step": 4880
},
{
"epoch": 15.92,
"learning_rate": 0.001,
"loss": 2.2141,
"step": 4890
},
{
"epoch": 15.95,
"learning_rate": 0.001,
"loss": 2.1724,
"step": 4900
},
{
"epoch": 15.98,
"learning_rate": 0.001,
"loss": 2.2171,
"step": 4910
},
{
"epoch": 16.02,
"learning_rate": 0.001,
"loss": 2.1489,
"step": 4920
},
{
"epoch": 16.05,
"learning_rate": 0.001,
"loss": 2.1308,
"step": 4930
},
{
"epoch": 16.08,
"learning_rate": 0.001,
"loss": 2.1169,
"step": 4940
},
{
"epoch": 16.12,
"learning_rate": 0.001,
"loss": 2.0744,
"step": 4950
},
{
"epoch": 16.15,
"learning_rate": 0.001,
"loss": 2.1527,
"step": 4960
},
{
"epoch": 16.18,
"learning_rate": 0.001,
"loss": 2.1514,
"step": 4970
},
{
"epoch": 16.21,
"learning_rate": 0.001,
"loss": 2.2027,
"step": 4980
},
{
"epoch": 16.25,
"learning_rate": 0.001,
"loss": 2.1814,
"step": 4990
},
{
"epoch": 16.28,
"learning_rate": 0.001,
"loss": 2.1292,
"step": 5000
},
{
"epoch": 16.28,
"eval_bp": 0.9822754617987439,
"eval_counts": [
27961,
19662,
14591,
11071
],
"eval_loss": 1.4380427598953247,
"eval_precisions": [
68.7813637705402,
52.87047245153137,
43.253097764866304,
36.513852242744065
],
"eval_ref_len": 41379,
"eval_runtime": 88.9695,
"eval_samples_per_second": 38.923,
"eval_score": 48.08648210020106,
"eval_steps_per_second": 38.923,
"eval_sys_len": 40652,
"eval_totals": [
40652,
37189,
33734,
30320
],
"step": 5000
},
{
"epoch": 16.31,
"learning_rate": 0.001,
"loss": 2.185,
"step": 5010
},
{
"epoch": 16.34,
"learning_rate": 0.001,
"loss": 2.1878,
"step": 5020
},
{
"epoch": 16.38,
"learning_rate": 0.001,
"loss": 2.1767,
"step": 5030
},
{
"epoch": 16.41,
"learning_rate": 0.001,
"loss": 2.011,
"step": 5040
},
{
"epoch": 16.44,
"learning_rate": 0.001,
"loss": 2.0795,
"step": 5050
},
{
"epoch": 16.47,
"learning_rate": 0.001,
"loss": 2.2293,
"step": 5060
},
{
"epoch": 16.51,
"learning_rate": 0.001,
"loss": 2.1918,
"step": 5070
},
{
"epoch": 16.54,
"learning_rate": 0.001,
"loss": 2.1273,
"step": 5080
},
{
"epoch": 16.57,
"learning_rate": 0.001,
"loss": 2.1646,
"step": 5090
},
{
"epoch": 16.6,
"learning_rate": 0.001,
"loss": 2.1413,
"step": 5100
},
{
"epoch": 16.64,
"learning_rate": 0.001,
"loss": 2.0892,
"step": 5110
},
{
"epoch": 16.67,
"learning_rate": 0.001,
"loss": 2.1659,
"step": 5120
},
{
"epoch": 16.7,
"learning_rate": 0.001,
"loss": 2.1922,
"step": 5130
},
{
"epoch": 16.73,
"learning_rate": 0.001,
"loss": 2.2153,
"step": 5140
},
{
"epoch": 16.77,
"learning_rate": 0.001,
"loss": 2.1297,
"step": 5150
},
{
"epoch": 16.8,
"learning_rate": 0.001,
"loss": 2.2311,
"step": 5160
},
{
"epoch": 16.83,
"learning_rate": 0.001,
"loss": 2.1413,
"step": 5170
},
{
"epoch": 16.86,
"learning_rate": 0.001,
"loss": 2.1689,
"step": 5180
},
{
"epoch": 16.9,
"learning_rate": 0.001,
"loss": 2.1698,
"step": 5190
},
{
"epoch": 16.93,
"learning_rate": 0.001,
"loss": 2.0922,
"step": 5200
},
{
"epoch": 16.93,
"eval_bp": 0.9634606403593156,
"eval_counts": [
27055,
18947,
14032,
10647
],
"eval_loss": 1.420142650604248,
"eval_precisions": [
67.8172156213967,
52.00790535532925,
42.54699818071558,
36.002434653230985
],
"eval_ref_len": 41379,
"eval_runtime": 88.5384,
"eval_samples_per_second": 39.113,
"eval_score": 46.450084211438956,
"eval_steps_per_second": 39.113,
"eval_sys_len": 39894,
"eval_totals": [
39894,
36431,
32980,
29573
],
"step": 5200
},
{
"epoch": 16.96,
"learning_rate": 0.001,
"loss": 2.212,
"step": 5210
},
{
"epoch": 16.99,
"learning_rate": 0.001,
"loss": 2.1306,
"step": 5220
},
{
"epoch": 17.03,
"learning_rate": 0.001,
"loss": 2.0919,
"step": 5230
},
{
"epoch": 17.06,
"learning_rate": 0.001,
"loss": 1.977,
"step": 5240
},
{
"epoch": 17.09,
"learning_rate": 0.001,
"loss": 2.0461,
"step": 5250
},
{
"epoch": 17.12,
"learning_rate": 0.001,
"loss": 2.1332,
"step": 5260
},
{
"epoch": 17.16,
"learning_rate": 0.001,
"loss": 2.116,
"step": 5270
},
{
"epoch": 17.19,
"learning_rate": 0.001,
"loss": 2.0639,
"step": 5280
},
{
"epoch": 17.22,
"learning_rate": 0.001,
"loss": 2.0844,
"step": 5290
},
{
"epoch": 17.25,
"learning_rate": 0.001,
"loss": 2.0423,
"step": 5300
},
{
"epoch": 17.29,
"learning_rate": 0.001,
"loss": 2.0831,
"step": 5310
},
{
"epoch": 17.32,
"learning_rate": 0.001,
"loss": 2.0964,
"step": 5320
},
{
"epoch": 17.35,
"learning_rate": 0.001,
"loss": 2.0526,
"step": 5330
},
{
"epoch": 17.38,
"learning_rate": 0.001,
"loss": 2.0584,
"step": 5340
},
{
"epoch": 17.42,
"learning_rate": 0.001,
"loss": 2.1096,
"step": 5350
},
{
"epoch": 17.45,
"learning_rate": 0.001,
"loss": 2.0576,
"step": 5360
},
{
"epoch": 17.48,
"learning_rate": 0.001,
"loss": 2.0952,
"step": 5370
},
{
"epoch": 17.51,
"learning_rate": 0.001,
"loss": 2.0561,
"step": 5380
},
{
"epoch": 17.55,
"learning_rate": 0.001,
"loss": 2.1358,
"step": 5390
},
{
"epoch": 17.58,
"learning_rate": 0.001,
"loss": 2.066,
"step": 5400
},
{
"epoch": 17.58,
"eval_bp": 1.0,
"eval_counts": [
27996,
19728,
14711,
11223
],
"eval_loss": 1.4336364269256592,
"eval_precisions": [
66.97607655502392,
51.45942562015807,
42.17240489636786,
35.660269445856635
],
"eval_ref_len": 41379,
"eval_runtime": 98.8235,
"eval_samples_per_second": 35.042,
"eval_score": 47.71439333194693,
"eval_steps_per_second": 35.042,
"eval_sys_len": 41800,
"eval_totals": [
41800,
38337,
34883,
31472
],
"step": 5400
},
{
"epoch": 17.61,
"learning_rate": 0.001,
"loss": 2.1324,
"step": 5410
},
{
"epoch": 17.65,
"learning_rate": 0.001,
"loss": 2.1562,
"step": 5420
},
{
"epoch": 17.68,
"learning_rate": 0.001,
"loss": 2.064,
"step": 5430
},
{
"epoch": 17.71,
"learning_rate": 0.001,
"loss": 2.1753,
"step": 5440
},
{
"epoch": 17.74,
"learning_rate": 0.001,
"loss": 2.1545,
"step": 5450
},
{
"epoch": 17.78,
"learning_rate": 0.001,
"loss": 2.1808,
"step": 5460
},
{
"epoch": 17.81,
"learning_rate": 0.001,
"loss": 1.9854,
"step": 5470
},
{
"epoch": 17.84,
"learning_rate": 0.001,
"loss": 2.1477,
"step": 5480
},
{
"epoch": 17.87,
"learning_rate": 0.001,
"loss": 2.147,
"step": 5490
},
{
"epoch": 17.91,
"learning_rate": 0.001,
"loss": 2.1071,
"step": 5500
},
{
"epoch": 17.94,
"learning_rate": 0.001,
"loss": 2.1168,
"step": 5510
},
{
"epoch": 17.97,
"learning_rate": 0.001,
"loss": 2.0857,
"step": 5520
},
{
"epoch": 18.0,
"learning_rate": 0.001,
"loss": 2.1181,
"step": 5530
},
{
"epoch": 18.04,
"learning_rate": 0.001,
"loss": 2.0222,
"step": 5540
},
{
"epoch": 18.07,
"learning_rate": 0.001,
"loss": 1.9394,
"step": 5550
},
{
"epoch": 18.1,
"learning_rate": 0.001,
"loss": 2.0617,
"step": 5560
},
{
"epoch": 18.13,
"learning_rate": 0.001,
"loss": 1.9781,
"step": 5570
},
{
"epoch": 18.17,
"learning_rate": 0.001,
"loss": 1.9584,
"step": 5580
},
{
"epoch": 18.2,
"learning_rate": 0.001,
"loss": 1.9962,
"step": 5590
},
{
"epoch": 18.23,
"learning_rate": 0.001,
"loss": 1.9179,
"step": 5600
},
{
"epoch": 18.23,
"eval_bp": 0.9862765874121399,
"eval_counts": [
28447,
20181,
15105,
11544
],
"eval_loss": 1.4454234838485718,
"eval_precisions": [
69.6974151659929,
54.02923538230885,
44.560151041359376,
37.866561700452664
],
"eval_ref_len": 41379,
"eval_runtime": 80.4217,
"eval_samples_per_second": 43.06,
"eval_score": 49.51773016565914,
"eval_steps_per_second": 43.06,
"eval_sys_len": 40815,
"eval_totals": [
40815,
37352,
33898,
30486
],
"step": 5600
}
],
"max_steps": 76750,
"num_train_epochs": 250,
"total_flos": 5.39878684251648e+16,
"trial_name": null,
"trial_params": null
}