unispeech-sat-base-timit-ft / trainer_state.json
patrickvonplaten's picture
End of training
4971d46
raw
history blame
41.7 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 20.0,
"global_step": 2900,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.07,
"learning_rate": 8.000000000000001e-07,
"loss": 9.5948,
"step": 10
},
{
"epoch": 0.14,
"learning_rate": 1.8e-06,
"loss": 8.4287,
"step": 20
},
{
"epoch": 0.21,
"learning_rate": 2.8000000000000003e-06,
"loss": 7.8808,
"step": 30
},
{
"epoch": 0.28,
"learning_rate": 3.7e-06,
"loss": 8.5345,
"step": 40
},
{
"epoch": 0.34,
"learning_rate": 4.7e-06,
"loss": 7.885,
"step": 50
},
{
"epoch": 0.41,
"learning_rate": 5.7000000000000005e-06,
"loss": 7.0681,
"step": 60
},
{
"epoch": 0.48,
"learning_rate": 6.700000000000001e-06,
"loss": 5.7587,
"step": 70
},
{
"epoch": 0.55,
"learning_rate": 7.6e-06,
"loss": 5.1226,
"step": 80
},
{
"epoch": 0.62,
"learning_rate": 8.599999999999999e-06,
"loss": 3.6571,
"step": 90
},
{
"epoch": 0.69,
"learning_rate": 9.600000000000001e-06,
"loss": 3.2582,
"step": 100
},
{
"epoch": 0.69,
"eval_loss": 3.1650655269622803,
"eval_runtime": 55.5137,
"eval_samples_per_second": 30.263,
"eval_steps_per_second": 30.263,
"eval_wer": 1.0,
"step": 100
},
{
"epoch": 0.76,
"learning_rate": 1.06e-05,
"loss": 3.1539,
"step": 110
},
{
"epoch": 0.83,
"learning_rate": 1.16e-05,
"loss": 3.0631,
"step": 120
},
{
"epoch": 0.9,
"learning_rate": 1.2600000000000001e-05,
"loss": 3.0257,
"step": 130
},
{
"epoch": 0.97,
"learning_rate": 1.3600000000000002e-05,
"loss": 3.0019,
"step": 140
},
{
"epoch": 1.03,
"learning_rate": 1.4599999999999999e-05,
"loss": 3.0388,
"step": 150
},
{
"epoch": 1.1,
"learning_rate": 1.56e-05,
"loss": 2.9952,
"step": 160
},
{
"epoch": 1.17,
"learning_rate": 1.66e-05,
"loss": 2.9761,
"step": 170
},
{
"epoch": 1.24,
"learning_rate": 1.76e-05,
"loss": 2.99,
"step": 180
},
{
"epoch": 1.31,
"learning_rate": 1.86e-05,
"loss": 2.9804,
"step": 190
},
{
"epoch": 1.38,
"learning_rate": 1.9600000000000002e-05,
"loss": 2.9542,
"step": 200
},
{
"epoch": 1.38,
"eval_loss": 2.9566688537597656,
"eval_runtime": 55.4569,
"eval_samples_per_second": 30.294,
"eval_steps_per_second": 30.294,
"eval_wer": 1.0,
"step": 200
},
{
"epoch": 1.45,
"learning_rate": 2.06e-05,
"loss": 2.9482,
"step": 210
},
{
"epoch": 1.52,
"learning_rate": 2.16e-05,
"loss": 2.9609,
"step": 220
},
{
"epoch": 1.59,
"learning_rate": 2.26e-05,
"loss": 2.9481,
"step": 230
},
{
"epoch": 1.66,
"learning_rate": 2.36e-05,
"loss": 2.9313,
"step": 240
},
{
"epoch": 1.72,
"learning_rate": 2.46e-05,
"loss": 2.9242,
"step": 250
},
{
"epoch": 1.79,
"learning_rate": 2.5600000000000002e-05,
"loss": 2.9556,
"step": 260
},
{
"epoch": 1.86,
"learning_rate": 2.6600000000000003e-05,
"loss": 2.9124,
"step": 270
},
{
"epoch": 1.93,
"learning_rate": 2.7600000000000003e-05,
"loss": 2.9213,
"step": 280
},
{
"epoch": 2.0,
"learning_rate": 2.86e-05,
"loss": 2.9496,
"step": 290
},
{
"epoch": 2.07,
"learning_rate": 2.96e-05,
"loss": 2.9656,
"step": 300
},
{
"epoch": 2.07,
"eval_loss": 2.919543504714966,
"eval_runtime": 55.4422,
"eval_samples_per_second": 30.302,
"eval_steps_per_second": 30.302,
"eval_wer": 1.0,
"step": 300
},
{
"epoch": 2.14,
"learning_rate": 3.06e-05,
"loss": 2.8978,
"step": 310
},
{
"epoch": 2.21,
"learning_rate": 3.16e-05,
"loss": 2.8973,
"step": 320
},
{
"epoch": 2.28,
"learning_rate": 3.26e-05,
"loss": 2.9267,
"step": 330
},
{
"epoch": 2.34,
"learning_rate": 3.3600000000000004e-05,
"loss": 2.894,
"step": 340
},
{
"epoch": 2.41,
"learning_rate": 3.46e-05,
"loss": 2.8864,
"step": 350
},
{
"epoch": 2.48,
"learning_rate": 3.56e-05,
"loss": 2.8939,
"step": 360
},
{
"epoch": 2.55,
"learning_rate": 3.66e-05,
"loss": 2.9002,
"step": 370
},
{
"epoch": 2.62,
"learning_rate": 3.76e-05,
"loss": 2.8699,
"step": 380
},
{
"epoch": 2.69,
"learning_rate": 3.86e-05,
"loss": 2.8659,
"step": 390
},
{
"epoch": 2.76,
"learning_rate": 3.960000000000001e-05,
"loss": 2.8946,
"step": 400
},
{
"epoch": 2.76,
"eval_loss": 2.8640990257263184,
"eval_runtime": 55.3518,
"eval_samples_per_second": 30.351,
"eval_steps_per_second": 30.351,
"eval_wer": 1.0,
"step": 400
},
{
"epoch": 2.83,
"learning_rate": 4.0600000000000004e-05,
"loss": 2.8589,
"step": 410
},
{
"epoch": 2.9,
"learning_rate": 4.16e-05,
"loss": 2.8074,
"step": 420
},
{
"epoch": 2.97,
"learning_rate": 4.26e-05,
"loss": 2.7744,
"step": 430
},
{
"epoch": 3.03,
"learning_rate": 4.36e-05,
"loss": 2.8088,
"step": 440
},
{
"epoch": 3.1,
"learning_rate": 4.46e-05,
"loss": 2.6376,
"step": 450
},
{
"epoch": 3.17,
"learning_rate": 4.5600000000000004e-05,
"loss": 2.5122,
"step": 460
},
{
"epoch": 3.24,
"learning_rate": 4.660000000000001e-05,
"loss": 2.4022,
"step": 470
},
{
"epoch": 3.31,
"learning_rate": 4.76e-05,
"loss": 2.3265,
"step": 480
},
{
"epoch": 3.38,
"learning_rate": 4.86e-05,
"loss": 2.0526,
"step": 490
},
{
"epoch": 3.45,
"learning_rate": 4.96e-05,
"loss": 1.9305,
"step": 500
},
{
"epoch": 3.45,
"eval_loss": 1.7680213451385498,
"eval_runtime": 55.2749,
"eval_samples_per_second": 30.394,
"eval_steps_per_second": 30.394,
"eval_wer": 1.0028943560057888,
"step": 500
},
{
"epoch": 3.52,
"learning_rate": 5.0600000000000003e-05,
"loss": 1.9207,
"step": 510
},
{
"epoch": 3.59,
"learning_rate": 5.16e-05,
"loss": 1.6255,
"step": 520
},
{
"epoch": 3.66,
"learning_rate": 5.2600000000000005e-05,
"loss": 1.4896,
"step": 530
},
{
"epoch": 3.72,
"learning_rate": 5.360000000000001e-05,
"loss": 1.5529,
"step": 540
},
{
"epoch": 3.79,
"learning_rate": 5.4600000000000006e-05,
"loss": 1.5405,
"step": 550
},
{
"epoch": 3.86,
"learning_rate": 5.560000000000001e-05,
"loss": 1.2831,
"step": 560
},
{
"epoch": 3.93,
"learning_rate": 5.66e-05,
"loss": 1.2698,
"step": 570
},
{
"epoch": 4.0,
"learning_rate": 5.76e-05,
"loss": 1.378,
"step": 580
},
{
"epoch": 4.07,
"learning_rate": 5.86e-05,
"loss": 1.2493,
"step": 590
},
{
"epoch": 4.14,
"learning_rate": 5.96e-05,
"loss": 1.0134,
"step": 600
},
{
"epoch": 4.14,
"eval_loss": 1.0184355974197388,
"eval_runtime": 55.9345,
"eval_samples_per_second": 30.035,
"eval_steps_per_second": 30.035,
"eval_wer": 0.6941630487216595,
"step": 600
},
{
"epoch": 4.21,
"learning_rate": 6.06e-05,
"loss": 1.0426,
"step": 610
},
{
"epoch": 4.28,
"learning_rate": 6.16e-05,
"loss": 1.2224,
"step": 620
},
{
"epoch": 4.34,
"learning_rate": 6.26e-05,
"loss": 0.9292,
"step": 630
},
{
"epoch": 4.41,
"learning_rate": 6.36e-05,
"loss": 0.9431,
"step": 640
},
{
"epoch": 4.48,
"learning_rate": 6.460000000000001e-05,
"loss": 1.0235,
"step": 650
},
{
"epoch": 4.55,
"learning_rate": 6.560000000000001e-05,
"loss": 1.0563,
"step": 660
},
{
"epoch": 4.62,
"learning_rate": 6.66e-05,
"loss": 0.8267,
"step": 670
},
{
"epoch": 4.69,
"learning_rate": 6.76e-05,
"loss": 0.9135,
"step": 680
},
{
"epoch": 4.76,
"learning_rate": 6.860000000000001e-05,
"loss": 1.0406,
"step": 690
},
{
"epoch": 4.83,
"learning_rate": 6.96e-05,
"loss": 0.8355,
"step": 700
},
{
"epoch": 4.83,
"eval_loss": 0.7769410610198975,
"eval_runtime": 55.6343,
"eval_samples_per_second": 30.197,
"eval_steps_per_second": 30.197,
"eval_wer": 0.6080215009303287,
"step": 700
},
{
"epoch": 4.9,
"learning_rate": 7.06e-05,
"loss": 0.7333,
"step": 710
},
{
"epoch": 4.97,
"learning_rate": 7.16e-05,
"loss": 0.8576,
"step": 720
},
{
"epoch": 5.03,
"learning_rate": 7.26e-05,
"loss": 0.9464,
"step": 730
},
{
"epoch": 5.1,
"learning_rate": 7.36e-05,
"loss": 0.6546,
"step": 740
},
{
"epoch": 5.17,
"learning_rate": 7.46e-05,
"loss": 0.6878,
"step": 750
},
{
"epoch": 5.24,
"learning_rate": 7.560000000000001e-05,
"loss": 0.8312,
"step": 760
},
{
"epoch": 5.31,
"learning_rate": 7.66e-05,
"loss": 0.7914,
"step": 770
},
{
"epoch": 5.38,
"learning_rate": 7.76e-05,
"loss": 0.625,
"step": 780
},
{
"epoch": 5.45,
"learning_rate": 7.860000000000001e-05,
"loss": 0.671,
"step": 790
},
{
"epoch": 5.52,
"learning_rate": 7.960000000000001e-05,
"loss": 0.8724,
"step": 800
},
{
"epoch": 5.52,
"eval_loss": 0.7182188034057617,
"eval_runtime": 55.6008,
"eval_samples_per_second": 30.215,
"eval_steps_per_second": 30.215,
"eval_wer": 0.6035421404451795,
"step": 800
},
{
"epoch": 5.59,
"learning_rate": 8.060000000000001e-05,
"loss": 0.6512,
"step": 810
},
{
"epoch": 5.66,
"learning_rate": 8.16e-05,
"loss": 0.5989,
"step": 820
},
{
"epoch": 5.72,
"learning_rate": 8.26e-05,
"loss": 0.7179,
"step": 830
},
{
"epoch": 5.79,
"learning_rate": 8.36e-05,
"loss": 0.8365,
"step": 840
},
{
"epoch": 5.86,
"learning_rate": 8.46e-05,
"loss": 0.5486,
"step": 850
},
{
"epoch": 5.93,
"learning_rate": 8.560000000000001e-05,
"loss": 0.6065,
"step": 860
},
{
"epoch": 6.0,
"learning_rate": 8.66e-05,
"loss": 0.7953,
"step": 870
},
{
"epoch": 6.07,
"learning_rate": 8.76e-05,
"loss": 0.5734,
"step": 880
},
{
"epoch": 6.14,
"learning_rate": 8.86e-05,
"loss": 0.4909,
"step": 890
},
{
"epoch": 6.21,
"learning_rate": 8.960000000000001e-05,
"loss": 0.5619,
"step": 900
},
{
"epoch": 6.21,
"eval_loss": 0.6822848916053772,
"eval_runtime": 56.1503,
"eval_samples_per_second": 29.92,
"eval_steps_per_second": 29.92,
"eval_wer": 0.5406243539383916,
"step": 900
},
{
"epoch": 6.28,
"learning_rate": 9.06e-05,
"loss": 0.7535,
"step": 910
},
{
"epoch": 6.34,
"learning_rate": 9.16e-05,
"loss": 0.5114,
"step": 920
},
{
"epoch": 6.41,
"learning_rate": 9.260000000000001e-05,
"loss": 0.5154,
"step": 930
},
{
"epoch": 6.48,
"learning_rate": 9.360000000000001e-05,
"loss": 0.6215,
"step": 940
},
{
"epoch": 6.55,
"learning_rate": 9.46e-05,
"loss": 0.6334,
"step": 950
},
{
"epoch": 6.62,
"learning_rate": 9.56e-05,
"loss": 0.4418,
"step": 960
},
{
"epoch": 6.69,
"learning_rate": 9.66e-05,
"loss": 0.4978,
"step": 970
},
{
"epoch": 6.76,
"learning_rate": 9.76e-05,
"loss": 0.7198,
"step": 980
},
{
"epoch": 6.83,
"learning_rate": 9.86e-05,
"loss": 0.5381,
"step": 990
},
{
"epoch": 6.9,
"learning_rate": 9.960000000000001e-05,
"loss": 0.4247,
"step": 1000
},
{
"epoch": 6.9,
"eval_loss": 0.6278514862060547,
"eval_runtime": 55.8229,
"eval_samples_per_second": 30.095,
"eval_steps_per_second": 30.095,
"eval_wer": 0.5237406105712907,
"step": 1000
},
{
"epoch": 6.97,
"learning_rate": 9.968421052631578e-05,
"loss": 0.5439,
"step": 1010
},
{
"epoch": 7.03,
"learning_rate": 9.915789473684211e-05,
"loss": 0.64,
"step": 1020
},
{
"epoch": 7.1,
"learning_rate": 9.863157894736842e-05,
"loss": 0.3949,
"step": 1030
},
{
"epoch": 7.17,
"learning_rate": 9.810526315789475e-05,
"loss": 0.4053,
"step": 1040
},
{
"epoch": 7.24,
"learning_rate": 9.757894736842106e-05,
"loss": 0.5467,
"step": 1050
},
{
"epoch": 7.31,
"learning_rate": 9.705263157894738e-05,
"loss": 0.5064,
"step": 1060
},
{
"epoch": 7.38,
"learning_rate": 9.652631578947369e-05,
"loss": 0.3485,
"step": 1070
},
{
"epoch": 7.45,
"learning_rate": 9.6e-05,
"loss": 0.4358,
"step": 1080
},
{
"epoch": 7.52,
"learning_rate": 9.547368421052632e-05,
"loss": 0.6029,
"step": 1090
},
{
"epoch": 7.59,
"learning_rate": 9.494736842105264e-05,
"loss": 0.4257,
"step": 1100
},
{
"epoch": 7.59,
"eval_loss": 0.6055994033813477,
"eval_runtime": 55.3764,
"eval_samples_per_second": 30.338,
"eval_steps_per_second": 30.338,
"eval_wer": 0.5000344566191165,
"step": 1100
},
{
"epoch": 7.66,
"learning_rate": 9.442105263157895e-05,
"loss": 0.3649,
"step": 1110
},
{
"epoch": 7.72,
"learning_rate": 9.389473684210527e-05,
"loss": 0.4477,
"step": 1120
},
{
"epoch": 7.79,
"learning_rate": 9.336842105263158e-05,
"loss": 0.5411,
"step": 1130
},
{
"epoch": 7.86,
"learning_rate": 9.28421052631579e-05,
"loss": 0.336,
"step": 1140
},
{
"epoch": 7.93,
"learning_rate": 9.231578947368421e-05,
"loss": 0.3939,
"step": 1150
},
{
"epoch": 8.0,
"learning_rate": 9.178947368421052e-05,
"loss": 0.5429,
"step": 1160
},
{
"epoch": 8.07,
"learning_rate": 9.126315789473685e-05,
"loss": 0.3765,
"step": 1170
},
{
"epoch": 8.14,
"learning_rate": 9.073684210526316e-05,
"loss": 0.2774,
"step": 1180
},
{
"epoch": 8.21,
"learning_rate": 9.021052631578947e-05,
"loss": 0.3571,
"step": 1190
},
{
"epoch": 8.28,
"learning_rate": 8.96842105263158e-05,
"loss": 0.5007,
"step": 1200
},
{
"epoch": 8.28,
"eval_loss": 0.5869982242584229,
"eval_runtime": 55.5221,
"eval_samples_per_second": 30.258,
"eval_steps_per_second": 30.258,
"eval_wer": 0.4918337812693819,
"step": 1200
},
{
"epoch": 8.34,
"learning_rate": 8.915789473684211e-05,
"loss": 0.3043,
"step": 1210
},
{
"epoch": 8.41,
"learning_rate": 8.863157894736842e-05,
"loss": 0.3133,
"step": 1220
},
{
"epoch": 8.48,
"learning_rate": 8.810526315789475e-05,
"loss": 0.4263,
"step": 1230
},
{
"epoch": 8.55,
"learning_rate": 8.757894736842106e-05,
"loss": 0.4459,
"step": 1240
},
{
"epoch": 8.62,
"learning_rate": 8.705263157894738e-05,
"loss": 0.2818,
"step": 1250
},
{
"epoch": 8.69,
"learning_rate": 8.652631578947369e-05,
"loss": 0.3313,
"step": 1260
},
{
"epoch": 8.76,
"learning_rate": 8.6e-05,
"loss": 0.4947,
"step": 1270
},
{
"epoch": 8.83,
"learning_rate": 8.547368421052632e-05,
"loss": 0.323,
"step": 1280
},
{
"epoch": 8.9,
"learning_rate": 8.494736842105263e-05,
"loss": 0.2867,
"step": 1290
},
{
"epoch": 8.97,
"learning_rate": 8.442105263157896e-05,
"loss": 0.3854,
"step": 1300
},
{
"epoch": 8.97,
"eval_loss": 0.6200077533721924,
"eval_runtime": 55.2433,
"eval_samples_per_second": 30.411,
"eval_steps_per_second": 30.411,
"eval_wer": 0.4803941837226931,
"step": 1300
},
{
"epoch": 9.03,
"learning_rate": 8.389473684210527e-05,
"loss": 0.4356,
"step": 1310
},
{
"epoch": 9.1,
"learning_rate": 8.336842105263158e-05,
"loss": 0.2332,
"step": 1320
},
{
"epoch": 9.17,
"learning_rate": 8.28421052631579e-05,
"loss": 0.2656,
"step": 1330
},
{
"epoch": 9.24,
"learning_rate": 8.231578947368421e-05,
"loss": 0.3761,
"step": 1340
},
{
"epoch": 9.31,
"learning_rate": 8.178947368421052e-05,
"loss": 0.3593,
"step": 1350
},
{
"epoch": 9.38,
"learning_rate": 8.126315789473685e-05,
"loss": 0.2388,
"step": 1360
},
{
"epoch": 9.45,
"learning_rate": 8.073684210526316e-05,
"loss": 0.2869,
"step": 1370
},
{
"epoch": 9.52,
"learning_rate": 8.021052631578949e-05,
"loss": 0.4014,
"step": 1380
},
{
"epoch": 9.59,
"learning_rate": 7.96842105263158e-05,
"loss": 0.2585,
"step": 1390
},
{
"epoch": 9.66,
"learning_rate": 7.915789473684211e-05,
"loss": 0.264,
"step": 1400
},
{
"epoch": 9.66,
"eval_loss": 0.6030145883560181,
"eval_runtime": 55.5584,
"eval_samples_per_second": 30.238,
"eval_steps_per_second": 30.238,
"eval_wer": 0.459995865205706,
"step": 1400
},
{
"epoch": 9.72,
"learning_rate": 7.863157894736843e-05,
"loss": 0.3439,
"step": 1410
},
{
"epoch": 9.79,
"learning_rate": 7.810526315789474e-05,
"loss": 0.3895,
"step": 1420
},
{
"epoch": 9.86,
"learning_rate": 7.757894736842105e-05,
"loss": 0.2278,
"step": 1430
},
{
"epoch": 9.93,
"learning_rate": 7.705263157894738e-05,
"loss": 0.2772,
"step": 1440
},
{
"epoch": 10.0,
"learning_rate": 7.652631578947369e-05,
"loss": 0.3984,
"step": 1450
},
{
"epoch": 10.07,
"learning_rate": 7.6e-05,
"loss": 0.2493,
"step": 1460
},
{
"epoch": 10.14,
"learning_rate": 7.547368421052632e-05,
"loss": 0.2003,
"step": 1470
},
{
"epoch": 10.21,
"learning_rate": 7.494736842105263e-05,
"loss": 0.2734,
"step": 1480
},
{
"epoch": 10.28,
"learning_rate": 7.442105263157894e-05,
"loss": 0.3931,
"step": 1490
},
{
"epoch": 10.34,
"learning_rate": 7.389473684210527e-05,
"loss": 0.1989,
"step": 1500
},
{
"epoch": 10.34,
"eval_loss": 0.6049150228500366,
"eval_runtime": 55.7466,
"eval_samples_per_second": 30.136,
"eval_steps_per_second": 30.136,
"eval_wer": 0.45875542691751087,
"step": 1500
},
{
"epoch": 10.41,
"learning_rate": 7.336842105263158e-05,
"loss": 0.218,
"step": 1510
},
{
"epoch": 10.48,
"learning_rate": 7.28421052631579e-05,
"loss": 0.3049,
"step": 1520
},
{
"epoch": 10.55,
"learning_rate": 7.231578947368421e-05,
"loss": 0.3151,
"step": 1530
},
{
"epoch": 10.62,
"learning_rate": 7.178947368421054e-05,
"loss": 0.2026,
"step": 1540
},
{
"epoch": 10.69,
"learning_rate": 7.126315789473685e-05,
"loss": 0.2733,
"step": 1550
},
{
"epoch": 10.76,
"learning_rate": 7.073684210526316e-05,
"loss": 0.3534,
"step": 1560
},
{
"epoch": 10.83,
"learning_rate": 7.021052631578949e-05,
"loss": 0.2302,
"step": 1570
},
{
"epoch": 10.9,
"learning_rate": 6.96842105263158e-05,
"loss": 0.2019,
"step": 1580
},
{
"epoch": 10.97,
"learning_rate": 6.915789473684211e-05,
"loss": 0.2623,
"step": 1590
},
{
"epoch": 11.03,
"learning_rate": 6.863157894736843e-05,
"loss": 0.3196,
"step": 1600
},
{
"epoch": 11.03,
"eval_loss": 0.5946373343467712,
"eval_runtime": 55.8167,
"eval_samples_per_second": 30.099,
"eval_steps_per_second": 30.099,
"eval_wer": 0.45992695196747296,
"step": 1600
},
{
"epoch": 11.1,
"learning_rate": 6.810526315789474e-05,
"loss": 0.171,
"step": 1610
},
{
"epoch": 11.17,
"learning_rate": 6.757894736842105e-05,
"loss": 0.1797,
"step": 1620
},
{
"epoch": 11.24,
"learning_rate": 6.705263157894737e-05,
"loss": 0.3063,
"step": 1630
},
{
"epoch": 11.31,
"learning_rate": 6.652631578947369e-05,
"loss": 0.2461,
"step": 1640
},
{
"epoch": 11.38,
"learning_rate": 6.6e-05,
"loss": 0.1774,
"step": 1650
},
{
"epoch": 11.45,
"learning_rate": 6.547368421052632e-05,
"loss": 0.2203,
"step": 1660
},
{
"epoch": 11.52,
"learning_rate": 6.494736842105263e-05,
"loss": 0.3276,
"step": 1670
},
{
"epoch": 11.59,
"learning_rate": 6.442105263157894e-05,
"loss": 0.185,
"step": 1680
},
{
"epoch": 11.66,
"learning_rate": 6.389473684210526e-05,
"loss": 0.1844,
"step": 1690
},
{
"epoch": 11.72,
"learning_rate": 6.336842105263158e-05,
"loss": 0.2622,
"step": 1700
},
{
"epoch": 11.72,
"eval_loss": 0.6282255053520203,
"eval_runtime": 56.2007,
"eval_samples_per_second": 29.893,
"eval_steps_per_second": 29.893,
"eval_wer": 0.44221624974157536,
"step": 1700
},
{
"epoch": 11.79,
"learning_rate": 6.28421052631579e-05,
"loss": 0.2972,
"step": 1710
},
{
"epoch": 11.86,
"learning_rate": 6.231578947368422e-05,
"loss": 0.1421,
"step": 1720
},
{
"epoch": 11.93,
"learning_rate": 6.178947368421054e-05,
"loss": 0.234,
"step": 1730
},
{
"epoch": 12.0,
"learning_rate": 6.126315789473685e-05,
"loss": 0.3268,
"step": 1740
},
{
"epoch": 12.07,
"learning_rate": 6.073684210526316e-05,
"loss": 0.181,
"step": 1750
},
{
"epoch": 12.14,
"learning_rate": 6.0210526315789475e-05,
"loss": 0.1642,
"step": 1760
},
{
"epoch": 12.21,
"learning_rate": 5.968421052631579e-05,
"loss": 0.2088,
"step": 1770
},
{
"epoch": 12.28,
"learning_rate": 5.9157894736842114e-05,
"loss": 0.2914,
"step": 1780
},
{
"epoch": 12.34,
"learning_rate": 5.863157894736843e-05,
"loss": 0.1558,
"step": 1790
},
{
"epoch": 12.41,
"learning_rate": 5.810526315789474e-05,
"loss": 0.1697,
"step": 1800
},
{
"epoch": 12.41,
"eval_loss": 0.655932605266571,
"eval_runtime": 56.1286,
"eval_samples_per_second": 29.931,
"eval_steps_per_second": 29.931,
"eval_wer": 0.4413203776445455,
"step": 1800
},
{
"epoch": 12.48,
"learning_rate": 5.7578947368421053e-05,
"loss": 0.2814,
"step": 1810
},
{
"epoch": 12.55,
"learning_rate": 5.7052631578947366e-05,
"loss": 0.243,
"step": 1820
},
{
"epoch": 12.62,
"learning_rate": 5.652631578947368e-05,
"loss": 0.1452,
"step": 1830
},
{
"epoch": 12.69,
"learning_rate": 5.6000000000000006e-05,
"loss": 0.1819,
"step": 1840
},
{
"epoch": 12.76,
"learning_rate": 5.547368421052632e-05,
"loss": 0.2672,
"step": 1850
},
{
"epoch": 12.83,
"learning_rate": 5.494736842105264e-05,
"loss": 0.166,
"step": 1860
},
{
"epoch": 12.9,
"learning_rate": 5.442105263157895e-05,
"loss": 0.1595,
"step": 1870
},
{
"epoch": 12.97,
"learning_rate": 5.3894736842105265e-05,
"loss": 0.2145,
"step": 1880
},
{
"epoch": 13.03,
"learning_rate": 5.336842105263158e-05,
"loss": 0.2245,
"step": 1890
},
{
"epoch": 13.1,
"learning_rate": 5.284210526315789e-05,
"loss": 0.1464,
"step": 1900
},
{
"epoch": 13.1,
"eval_loss": 0.6348641514778137,
"eval_runtime": 55.8927,
"eval_samples_per_second": 30.058,
"eval_steps_per_second": 30.058,
"eval_wer": 0.4327751361036455,
"step": 1900
},
{
"epoch": 13.17,
"learning_rate": 5.231578947368422e-05,
"loss": 0.1798,
"step": 1910
},
{
"epoch": 13.24,
"learning_rate": 5.178947368421053e-05,
"loss": 0.2327,
"step": 1920
},
{
"epoch": 13.31,
"learning_rate": 5.126315789473684e-05,
"loss": 0.1854,
"step": 1930
},
{
"epoch": 13.38,
"learning_rate": 5.073684210526316e-05,
"loss": 0.1386,
"step": 1940
},
{
"epoch": 13.45,
"learning_rate": 5.0210526315789476e-05,
"loss": 0.1891,
"step": 1950
},
{
"epoch": 13.52,
"learning_rate": 4.9684210526315796e-05,
"loss": 0.2851,
"step": 1960
},
{
"epoch": 13.59,
"learning_rate": 4.915789473684211e-05,
"loss": 0.1592,
"step": 1970
},
{
"epoch": 13.66,
"learning_rate": 4.863157894736842e-05,
"loss": 0.1503,
"step": 1980
},
{
"epoch": 13.72,
"learning_rate": 4.8105263157894735e-05,
"loss": 0.232,
"step": 1990
},
{
"epoch": 13.79,
"learning_rate": 4.7578947368421054e-05,
"loss": 0.2277,
"step": 2000
},
{
"epoch": 13.79,
"eval_loss": 0.6132542490959167,
"eval_runtime": 56.1305,
"eval_samples_per_second": 29.93,
"eval_steps_per_second": 29.93,
"eval_wer": 0.4283646888567294,
"step": 2000
},
{
"epoch": 13.86,
"learning_rate": 4.705263157894737e-05,
"loss": 0.1354,
"step": 2010
},
{
"epoch": 13.93,
"learning_rate": 4.652631578947369e-05,
"loss": 0.1661,
"step": 2020
},
{
"epoch": 14.0,
"learning_rate": 4.600000000000001e-05,
"loss": 0.2663,
"step": 2030
},
{
"epoch": 14.07,
"learning_rate": 4.547368421052632e-05,
"loss": 0.1451,
"step": 2040
},
{
"epoch": 14.14,
"learning_rate": 4.494736842105263e-05,
"loss": 0.1315,
"step": 2050
},
{
"epoch": 14.21,
"learning_rate": 4.442105263157895e-05,
"loss": 0.1791,
"step": 2060
},
{
"epoch": 14.28,
"learning_rate": 4.3894736842105266e-05,
"loss": 0.2521,
"step": 2070
},
{
"epoch": 14.34,
"learning_rate": 4.336842105263158e-05,
"loss": 0.1194,
"step": 2080
},
{
"epoch": 14.41,
"learning_rate": 4.284210526315789e-05,
"loss": 0.1591,
"step": 2090
},
{
"epoch": 14.48,
"learning_rate": 4.231578947368421e-05,
"loss": 0.221,
"step": 2100
},
{
"epoch": 14.48,
"eval_loss": 0.6617235541343689,
"eval_runtime": 56.1176,
"eval_samples_per_second": 29.937,
"eval_steps_per_second": 29.937,
"eval_wer": 0.4218868444628213,
"step": 2100
},
{
"epoch": 14.55,
"learning_rate": 4.178947368421053e-05,
"loss": 0.2025,
"step": 2110
},
{
"epoch": 14.62,
"learning_rate": 4.1263157894736844e-05,
"loss": 0.1239,
"step": 2120
},
{
"epoch": 14.69,
"learning_rate": 4.0736842105263164e-05,
"loss": 0.1833,
"step": 2130
},
{
"epoch": 14.76,
"learning_rate": 4.021052631578948e-05,
"loss": 0.2398,
"step": 2140
},
{
"epoch": 14.83,
"learning_rate": 3.968421052631579e-05,
"loss": 0.1389,
"step": 2150
},
{
"epoch": 14.9,
"learning_rate": 3.91578947368421e-05,
"loss": 0.1291,
"step": 2160
},
{
"epoch": 14.97,
"learning_rate": 3.863157894736842e-05,
"loss": 0.1669,
"step": 2170
},
{
"epoch": 15.03,
"learning_rate": 3.8105263157894735e-05,
"loss": 0.2176,
"step": 2180
},
{
"epoch": 15.1,
"learning_rate": 3.7578947368421055e-05,
"loss": 0.1196,
"step": 2190
},
{
"epoch": 15.17,
"learning_rate": 3.7052631578947375e-05,
"loss": 0.1391,
"step": 2200
},
{
"epoch": 15.17,
"eval_loss": 0.6704933643341064,
"eval_runtime": 55.8352,
"eval_samples_per_second": 30.089,
"eval_steps_per_second": 30.089,
"eval_wer": 0.4234718489421818,
"step": 2200
},
{
"epoch": 15.24,
"learning_rate": 3.652631578947369e-05,
"loss": 0.2142,
"step": 2210
},
{
"epoch": 15.31,
"learning_rate": 3.6e-05,
"loss": 0.1519,
"step": 2220
},
{
"epoch": 15.38,
"learning_rate": 3.547368421052632e-05,
"loss": 0.1236,
"step": 2230
},
{
"epoch": 15.45,
"learning_rate": 3.4947368421052634e-05,
"loss": 0.1452,
"step": 2240
},
{
"epoch": 15.52,
"learning_rate": 3.442105263157895e-05,
"loss": 0.2405,
"step": 2250
},
{
"epoch": 15.59,
"learning_rate": 3.389473684210526e-05,
"loss": 0.1229,
"step": 2260
},
{
"epoch": 15.66,
"learning_rate": 3.336842105263158e-05,
"loss": 0.1357,
"step": 2270
},
{
"epoch": 15.72,
"learning_rate": 3.28421052631579e-05,
"loss": 0.2144,
"step": 2280
},
{
"epoch": 15.79,
"learning_rate": 3.231578947368421e-05,
"loss": 0.192,
"step": 2290
},
{
"epoch": 15.86,
"learning_rate": 3.178947368421053e-05,
"loss": 0.112,
"step": 2300
},
{
"epoch": 15.86,
"eval_loss": 0.6207090616226196,
"eval_runtime": 56.0432,
"eval_samples_per_second": 29.977,
"eval_steps_per_second": 29.977,
"eval_wer": 0.4218179312245882,
"step": 2300
},
{
"epoch": 15.93,
"learning_rate": 3.1263157894736845e-05,
"loss": 0.1391,
"step": 2310
},
{
"epoch": 16.0,
"learning_rate": 3.073684210526316e-05,
"loss": 0.2383,
"step": 2320
},
{
"epoch": 16.07,
"learning_rate": 3.021052631578947e-05,
"loss": 0.1233,
"step": 2330
},
{
"epoch": 16.14,
"learning_rate": 2.968421052631579e-05,
"loss": 0.1164,
"step": 2340
},
{
"epoch": 16.21,
"learning_rate": 2.9157894736842107e-05,
"loss": 0.1582,
"step": 2350
},
{
"epoch": 16.28,
"learning_rate": 2.863157894736842e-05,
"loss": 0.2145,
"step": 2360
},
{
"epoch": 16.34,
"learning_rate": 2.810526315789474e-05,
"loss": 0.1143,
"step": 2370
},
{
"epoch": 16.41,
"learning_rate": 2.7578947368421053e-05,
"loss": 0.1321,
"step": 2380
},
{
"epoch": 16.48,
"learning_rate": 2.705263157894737e-05,
"loss": 0.1853,
"step": 2390
},
{
"epoch": 16.55,
"learning_rate": 2.652631578947369e-05,
"loss": 0.1717,
"step": 2400
},
{
"epoch": 16.55,
"eval_loss": 0.6749277710914612,
"eval_runtime": 56.2422,
"eval_samples_per_second": 29.871,
"eval_steps_per_second": 29.871,
"eval_wer": 0.41837226931293503,
"step": 2400
},
{
"epoch": 16.62,
"learning_rate": 2.6000000000000002e-05,
"loss": 0.1022,
"step": 2410
},
{
"epoch": 16.69,
"learning_rate": 2.5473684210526315e-05,
"loss": 0.1444,
"step": 2420
},
{
"epoch": 16.76,
"learning_rate": 2.4947368421052635e-05,
"loss": 0.221,
"step": 2430
},
{
"epoch": 16.83,
"learning_rate": 2.4421052631578948e-05,
"loss": 0.1304,
"step": 2440
},
{
"epoch": 16.9,
"learning_rate": 2.3894736842105264e-05,
"loss": 0.1238,
"step": 2450
},
{
"epoch": 16.97,
"learning_rate": 2.336842105263158e-05,
"loss": 0.1517,
"step": 2460
},
{
"epoch": 17.03,
"learning_rate": 2.2842105263157897e-05,
"loss": 0.1952,
"step": 2470
},
{
"epoch": 17.1,
"learning_rate": 2.2315789473684213e-05,
"loss": 0.0988,
"step": 2480
},
{
"epoch": 17.17,
"learning_rate": 2.1789473684210526e-05,
"loss": 0.1344,
"step": 2490
},
{
"epoch": 17.24,
"learning_rate": 2.1263157894736842e-05,
"loss": 0.2081,
"step": 2500
},
{
"epoch": 17.24,
"eval_loss": 0.6755503416061401,
"eval_runtime": 56.1675,
"eval_samples_per_second": 29.911,
"eval_steps_per_second": 29.911,
"eval_wer": 0.4168561780718076,
"step": 2500
},
{
"epoch": 17.31,
"learning_rate": 2.073684210526316e-05,
"loss": 0.1461,
"step": 2510
},
{
"epoch": 17.38,
"learning_rate": 2.0210526315789475e-05,
"loss": 0.0932,
"step": 2520
},
{
"epoch": 17.45,
"learning_rate": 1.968421052631579e-05,
"loss": 0.1341,
"step": 2530
},
{
"epoch": 17.52,
"learning_rate": 1.9157894736842104e-05,
"loss": 0.1959,
"step": 2540
},
{
"epoch": 17.59,
"learning_rate": 1.863157894736842e-05,
"loss": 0.1091,
"step": 2550
},
{
"epoch": 17.66,
"learning_rate": 1.810526315789474e-05,
"loss": 0.1192,
"step": 2560
},
{
"epoch": 17.72,
"learning_rate": 1.7578947368421054e-05,
"loss": 0.1698,
"step": 2570
},
{
"epoch": 17.79,
"learning_rate": 1.705263157894737e-05,
"loss": 0.1667,
"step": 2580
},
{
"epoch": 17.86,
"learning_rate": 1.6526315789473683e-05,
"loss": 0.1047,
"step": 2590
},
{
"epoch": 17.93,
"learning_rate": 1.6000000000000003e-05,
"loss": 0.1244,
"step": 2600
},
{
"epoch": 17.93,
"eval_loss": 0.6749771237373352,
"eval_runtime": 56.0157,
"eval_samples_per_second": 29.992,
"eval_steps_per_second": 29.992,
"eval_wer": 0.4180966163600028,
"step": 2600
},
{
"epoch": 18.0,
"learning_rate": 1.5473684210526316e-05,
"loss": 0.1953,
"step": 2610
},
{
"epoch": 18.07,
"learning_rate": 1.4947368421052632e-05,
"loss": 0.1092,
"step": 2620
},
{
"epoch": 18.14,
"learning_rate": 1.4421052631578948e-05,
"loss": 0.1035,
"step": 2630
},
{
"epoch": 18.21,
"learning_rate": 1.3894736842105263e-05,
"loss": 0.1416,
"step": 2640
},
{
"epoch": 18.28,
"learning_rate": 1.336842105263158e-05,
"loss": 0.1985,
"step": 2650
},
{
"epoch": 18.34,
"learning_rate": 1.2842105263157894e-05,
"loss": 0.1042,
"step": 2660
},
{
"epoch": 18.41,
"learning_rate": 1.231578947368421e-05,
"loss": 0.1152,
"step": 2670
},
{
"epoch": 18.48,
"learning_rate": 1.1789473684210527e-05,
"loss": 0.1832,
"step": 2680
},
{
"epoch": 18.55,
"learning_rate": 1.1263157894736842e-05,
"loss": 0.1471,
"step": 2690
},
{
"epoch": 18.62,
"learning_rate": 1.073684210526316e-05,
"loss": 0.0978,
"step": 2700
},
{
"epoch": 18.62,
"eval_loss": 0.6500299572944641,
"eval_runtime": 56.1111,
"eval_samples_per_second": 29.941,
"eval_steps_per_second": 29.941,
"eval_wer": 0.41148094548962855,
"step": 2700
},
{
"epoch": 18.69,
"learning_rate": 1.0210526315789474e-05,
"loss": 0.145,
"step": 2710
},
{
"epoch": 18.76,
"learning_rate": 9.68421052631579e-06,
"loss": 0.1861,
"step": 2720
},
{
"epoch": 18.83,
"learning_rate": 9.157894736842105e-06,
"loss": 0.0956,
"step": 2730
},
{
"epoch": 18.9,
"learning_rate": 8.631578947368422e-06,
"loss": 0.121,
"step": 2740
},
{
"epoch": 18.97,
"learning_rate": 8.105263157894736e-06,
"loss": 0.1625,
"step": 2750
},
{
"epoch": 19.03,
"learning_rate": 7.578947368421054e-06,
"loss": 0.1652,
"step": 2760
},
{
"epoch": 19.1,
"learning_rate": 7.052631578947369e-06,
"loss": 0.0894,
"step": 2770
},
{
"epoch": 19.17,
"learning_rate": 6.526315789473685e-06,
"loss": 0.1145,
"step": 2780
},
{
"epoch": 19.24,
"learning_rate": 6e-06,
"loss": 0.1903,
"step": 2790
},
{
"epoch": 19.31,
"learning_rate": 5.4736842105263165e-06,
"loss": 0.128,
"step": 2800
},
{
"epoch": 19.31,
"eval_loss": 0.6750107407569885,
"eval_runtime": 55.7297,
"eval_samples_per_second": 30.146,
"eval_steps_per_second": 30.146,
"eval_wer": 0.4105850733925987,
"step": 2800
},
{
"epoch": 19.38,
"learning_rate": 4.947368421052632e-06,
"loss": 0.0957,
"step": 2810
},
{
"epoch": 19.45,
"learning_rate": 4.4210526315789476e-06,
"loss": 0.1292,
"step": 2820
},
{
"epoch": 19.52,
"learning_rate": 3.894736842105264e-06,
"loss": 0.1641,
"step": 2830
},
{
"epoch": 19.59,
"learning_rate": 3.368421052631579e-06,
"loss": 0.0968,
"step": 2840
},
{
"epoch": 19.66,
"learning_rate": 2.842105263157895e-06,
"loss": 0.1,
"step": 2850
},
{
"epoch": 19.72,
"learning_rate": 2.315789473684211e-06,
"loss": 0.1546,
"step": 2860
},
{
"epoch": 19.79,
"learning_rate": 1.7894736842105262e-06,
"loss": 0.1656,
"step": 2870
},
{
"epoch": 19.86,
"learning_rate": 1.2631578947368422e-06,
"loss": 0.1,
"step": 2880
},
{
"epoch": 19.93,
"learning_rate": 7.368421052631579e-07,
"loss": 0.1112,
"step": 2890
},
{
"epoch": 20.0,
"learning_rate": 2.105263157894737e-07,
"loss": 0.1791,
"step": 2900
},
{
"epoch": 20.0,
"eval_loss": 0.6711965203285217,
"eval_runtime": 56.2882,
"eval_samples_per_second": 29.846,
"eval_steps_per_second": 29.846,
"eval_wer": 0.41010268072496725,
"step": 2900
},
{
"epoch": 20.0,
"step": 2900,
"total_flos": 2.6571394113263534e+18,
"train_loss": 0.1187238040669211,
"train_runtime": 2793.6874,
"train_samples_per_second": 33.075,
"train_steps_per_second": 1.038
}
],
"max_steps": 2900,
"num_train_epochs": 20,
"total_flos": 2.6571394113263534e+18,
"trial_name": null,
"trial_params": null
}