t5_3b_epoch_3_qa / trainer_state.json
learn3r's picture
End of training
f3bbb1b
{
"best_metric": 0.2233106642961502,
"best_model_checkpoint": "/exports/eddie/scratch/s1970716/models/question_gen/t5_3b_epoch_3_qa/checkpoint-342",
"epoch": 2.9983561643835617,
"global_step": 1026,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.03,
"learning_rate": 9.902534113060428e-05,
"loss": 0.2782,
"step": 10
},
{
"epoch": 0.06,
"learning_rate": 9.805068226120859e-05,
"loss": 0.221,
"step": 20
},
{
"epoch": 0.09,
"learning_rate": 9.707602339181286e-05,
"loss": 0.2068,
"step": 30
},
{
"epoch": 0.12,
"learning_rate": 9.610136452241715e-05,
"loss": 0.2119,
"step": 40
},
{
"epoch": 0.15,
"learning_rate": 9.512670565302145e-05,
"loss": 0.2116,
"step": 50
},
{
"epoch": 0.18,
"learning_rate": 9.415204678362574e-05,
"loss": 0.2104,
"step": 60
},
{
"epoch": 0.2,
"learning_rate": 9.317738791423001e-05,
"loss": 0.2137,
"step": 70
},
{
"epoch": 0.23,
"learning_rate": 9.220272904483432e-05,
"loss": 0.2157,
"step": 80
},
{
"epoch": 0.26,
"learning_rate": 9.12280701754386e-05,
"loss": 0.2052,
"step": 90
},
{
"epoch": 0.29,
"learning_rate": 9.025341130604289e-05,
"loss": 0.2062,
"step": 100
},
{
"epoch": 0.32,
"learning_rate": 8.927875243664718e-05,
"loss": 0.2091,
"step": 110
},
{
"epoch": 0.35,
"learning_rate": 8.830409356725147e-05,
"loss": 0.2161,
"step": 120
},
{
"epoch": 0.38,
"learning_rate": 8.732943469785574e-05,
"loss": 0.2051,
"step": 130
},
{
"epoch": 0.41,
"learning_rate": 8.635477582846005e-05,
"loss": 0.2112,
"step": 140
},
{
"epoch": 0.44,
"learning_rate": 8.538011695906433e-05,
"loss": 0.1951,
"step": 150
},
{
"epoch": 0.47,
"learning_rate": 8.440545808966862e-05,
"loss": 0.2062,
"step": 160
},
{
"epoch": 0.5,
"learning_rate": 8.343079922027291e-05,
"loss": 0.2046,
"step": 170
},
{
"epoch": 0.53,
"learning_rate": 8.24561403508772e-05,
"loss": 0.1999,
"step": 180
},
{
"epoch": 0.56,
"learning_rate": 8.148148148148148e-05,
"loss": 0.2047,
"step": 190
},
{
"epoch": 0.58,
"learning_rate": 8.050682261208578e-05,
"loss": 0.2042,
"step": 200
},
{
"epoch": 0.61,
"learning_rate": 7.953216374269006e-05,
"loss": 0.1995,
"step": 210
},
{
"epoch": 0.64,
"learning_rate": 7.855750487329435e-05,
"loss": 0.1926,
"step": 220
},
{
"epoch": 0.67,
"learning_rate": 7.758284600389864e-05,
"loss": 0.204,
"step": 230
},
{
"epoch": 0.7,
"learning_rate": 7.660818713450293e-05,
"loss": 0.1925,
"step": 240
},
{
"epoch": 0.73,
"learning_rate": 7.563352826510721e-05,
"loss": 0.2017,
"step": 250
},
{
"epoch": 0.76,
"learning_rate": 7.465886939571151e-05,
"loss": 0.1984,
"step": 260
},
{
"epoch": 0.79,
"learning_rate": 7.368421052631579e-05,
"loss": 0.1941,
"step": 270
},
{
"epoch": 0.82,
"learning_rate": 7.270955165692008e-05,
"loss": 0.1932,
"step": 280
},
{
"epoch": 0.85,
"learning_rate": 7.173489278752437e-05,
"loss": 0.1979,
"step": 290
},
{
"epoch": 0.88,
"learning_rate": 7.076023391812866e-05,
"loss": 0.1994,
"step": 300
},
{
"epoch": 0.91,
"learning_rate": 6.978557504873294e-05,
"loss": 0.1935,
"step": 310
},
{
"epoch": 0.94,
"learning_rate": 6.881091617933724e-05,
"loss": 0.1951,
"step": 320
},
{
"epoch": 0.96,
"learning_rate": 6.783625730994152e-05,
"loss": 0.1964,
"step": 330
},
{
"epoch": 0.99,
"learning_rate": 6.686159844054581e-05,
"loss": 0.2003,
"step": 340
},
{
"epoch": 1.0,
"eval_loss": 0.2233106642961502,
"eval_runtime": 133.4707,
"eval_samples_per_second": 79.193,
"eval_steps_per_second": 9.905,
"step": 342
},
{
"epoch": 1.02,
"learning_rate": 6.58869395711501e-05,
"loss": 0.1463,
"step": 350
},
{
"epoch": 1.05,
"learning_rate": 6.49122807017544e-05,
"loss": 0.1331,
"step": 360
},
{
"epoch": 1.08,
"learning_rate": 6.393762183235867e-05,
"loss": 0.1401,
"step": 370
},
{
"epoch": 1.11,
"learning_rate": 6.296296296296296e-05,
"loss": 0.1359,
"step": 380
},
{
"epoch": 1.14,
"learning_rate": 6.198830409356725e-05,
"loss": 0.1372,
"step": 390
},
{
"epoch": 1.17,
"learning_rate": 6.101364522417154e-05,
"loss": 0.1361,
"step": 400
},
{
"epoch": 1.2,
"learning_rate": 6.003898635477583e-05,
"loss": 0.1408,
"step": 410
},
{
"epoch": 1.23,
"learning_rate": 5.9064327485380125e-05,
"loss": 0.1382,
"step": 420
},
{
"epoch": 1.26,
"learning_rate": 5.80896686159844e-05,
"loss": 0.1316,
"step": 430
},
{
"epoch": 1.29,
"learning_rate": 5.71150097465887e-05,
"loss": 0.1416,
"step": 440
},
{
"epoch": 1.32,
"learning_rate": 5.6140350877192984e-05,
"loss": 0.1371,
"step": 450
},
{
"epoch": 1.34,
"learning_rate": 5.5165692007797275e-05,
"loss": 0.1366,
"step": 460
},
{
"epoch": 1.37,
"learning_rate": 5.419103313840156e-05,
"loss": 0.1359,
"step": 470
},
{
"epoch": 1.4,
"learning_rate": 5.3216374269005856e-05,
"loss": 0.1395,
"step": 480
},
{
"epoch": 1.43,
"learning_rate": 5.2241715399610133e-05,
"loss": 0.1353,
"step": 490
},
{
"epoch": 1.46,
"learning_rate": 5.126705653021443e-05,
"loss": 0.1369,
"step": 500
},
{
"epoch": 1.49,
"learning_rate": 5.0292397660818715e-05,
"loss": 0.1296,
"step": 510
},
{
"epoch": 1.52,
"learning_rate": 4.9317738791423e-05,
"loss": 0.1323,
"step": 520
},
{
"epoch": 1.55,
"learning_rate": 4.834307992202729e-05,
"loss": 0.1363,
"step": 530
},
{
"epoch": 1.58,
"learning_rate": 4.736842105263158e-05,
"loss": 0.1492,
"step": 540
},
{
"epoch": 1.61,
"learning_rate": 4.6393762183235865e-05,
"loss": 0.1355,
"step": 550
},
{
"epoch": 1.64,
"learning_rate": 4.5419103313840156e-05,
"loss": 0.134,
"step": 560
},
{
"epoch": 1.67,
"learning_rate": 4.4444444444444447e-05,
"loss": 0.1385,
"step": 570
},
{
"epoch": 1.69,
"learning_rate": 4.346978557504873e-05,
"loss": 0.1301,
"step": 580
},
{
"epoch": 1.72,
"learning_rate": 4.249512670565302e-05,
"loss": 0.1442,
"step": 590
},
{
"epoch": 1.75,
"learning_rate": 4.152046783625731e-05,
"loss": 0.135,
"step": 600
},
{
"epoch": 1.78,
"learning_rate": 4.0545808966861596e-05,
"loss": 0.1278,
"step": 610
},
{
"epoch": 1.81,
"learning_rate": 3.957115009746589e-05,
"loss": 0.1389,
"step": 620
},
{
"epoch": 1.84,
"learning_rate": 3.859649122807018e-05,
"loss": 0.1386,
"step": 630
},
{
"epoch": 1.87,
"learning_rate": 3.762183235867446e-05,
"loss": 0.1313,
"step": 640
},
{
"epoch": 1.9,
"learning_rate": 3.664717348927875e-05,
"loss": 0.1325,
"step": 650
},
{
"epoch": 1.93,
"learning_rate": 3.5672514619883044e-05,
"loss": 0.1315,
"step": 660
},
{
"epoch": 1.96,
"learning_rate": 3.469785575048733e-05,
"loss": 0.1344,
"step": 670
},
{
"epoch": 1.99,
"learning_rate": 3.372319688109162e-05,
"loss": 0.1331,
"step": 680
},
{
"epoch": 2.0,
"eval_loss": 0.2396751344203949,
"eval_runtime": 133.3515,
"eval_samples_per_second": 79.264,
"eval_steps_per_second": 9.914,
"step": 684
},
{
"epoch": 2.02,
"learning_rate": 3.274853801169591e-05,
"loss": 0.1059,
"step": 690
},
{
"epoch": 2.05,
"learning_rate": 3.1773879142300193e-05,
"loss": 0.1006,
"step": 700
},
{
"epoch": 2.07,
"learning_rate": 3.0799220272904484e-05,
"loss": 0.0941,
"step": 710
},
{
"epoch": 2.1,
"learning_rate": 2.9824561403508772e-05,
"loss": 0.1033,
"step": 720
},
{
"epoch": 2.13,
"learning_rate": 2.884990253411306e-05,
"loss": 0.0978,
"step": 730
},
{
"epoch": 2.16,
"learning_rate": 2.787524366471735e-05,
"loss": 0.0983,
"step": 740
},
{
"epoch": 2.19,
"learning_rate": 2.6900584795321637e-05,
"loss": 0.0987,
"step": 750
},
{
"epoch": 2.22,
"learning_rate": 2.5925925925925925e-05,
"loss": 0.0985,
"step": 760
},
{
"epoch": 2.25,
"learning_rate": 2.4951267056530216e-05,
"loss": 0.0974,
"step": 770
},
{
"epoch": 2.28,
"learning_rate": 2.3976608187134503e-05,
"loss": 0.1001,
"step": 780
},
{
"epoch": 2.31,
"learning_rate": 2.300194931773879e-05,
"loss": 0.0961,
"step": 790
},
{
"epoch": 2.34,
"learning_rate": 2.2027290448343078e-05,
"loss": 0.0958,
"step": 800
},
{
"epoch": 2.37,
"learning_rate": 2.105263157894737e-05,
"loss": 0.0975,
"step": 810
},
{
"epoch": 2.4,
"learning_rate": 2.0077972709551656e-05,
"loss": 0.0955,
"step": 820
},
{
"epoch": 2.43,
"learning_rate": 1.9103313840155944e-05,
"loss": 0.0927,
"step": 830
},
{
"epoch": 2.45,
"learning_rate": 1.8128654970760235e-05,
"loss": 0.0958,
"step": 840
},
{
"epoch": 2.48,
"learning_rate": 1.7153996101364522e-05,
"loss": 0.0979,
"step": 850
},
{
"epoch": 2.51,
"learning_rate": 1.617933723196881e-05,
"loss": 0.1,
"step": 860
},
{
"epoch": 2.54,
"learning_rate": 1.5204678362573099e-05,
"loss": 0.0965,
"step": 870
},
{
"epoch": 2.57,
"learning_rate": 1.4230019493177388e-05,
"loss": 0.098,
"step": 880
},
{
"epoch": 2.6,
"learning_rate": 1.3255360623781677e-05,
"loss": 0.0981,
"step": 890
},
{
"epoch": 2.63,
"learning_rate": 1.2280701754385964e-05,
"loss": 0.1024,
"step": 900
},
{
"epoch": 2.66,
"learning_rate": 1.1306042884990253e-05,
"loss": 0.0962,
"step": 910
},
{
"epoch": 2.69,
"learning_rate": 1.0331384015594543e-05,
"loss": 0.0993,
"step": 920
},
{
"epoch": 2.72,
"learning_rate": 9.35672514619883e-06,
"loss": 0.0932,
"step": 930
},
{
"epoch": 2.75,
"learning_rate": 8.38206627680312e-06,
"loss": 0.0929,
"step": 940
},
{
"epoch": 2.78,
"learning_rate": 7.4074074074074075e-06,
"loss": 0.0881,
"step": 950
},
{
"epoch": 2.81,
"learning_rate": 6.432748538011696e-06,
"loss": 0.0923,
"step": 960
},
{
"epoch": 2.83,
"learning_rate": 5.458089668615984e-06,
"loss": 0.0935,
"step": 970
},
{
"epoch": 2.86,
"learning_rate": 4.483430799220273e-06,
"loss": 0.1015,
"step": 980
},
{
"epoch": 2.89,
"learning_rate": 3.5087719298245615e-06,
"loss": 0.0939,
"step": 990
},
{
"epoch": 2.92,
"learning_rate": 2.5341130604288498e-06,
"loss": 0.093,
"step": 1000
},
{
"epoch": 2.95,
"learning_rate": 1.5594541910331385e-06,
"loss": 0.0975,
"step": 1010
},
{
"epoch": 2.98,
"learning_rate": 5.847953216374269e-07,
"loss": 0.0971,
"step": 1020
},
{
"epoch": 3.0,
"eval_loss": 0.26710498332977295,
"eval_runtime": 133.9576,
"eval_samples_per_second": 78.906,
"eval_steps_per_second": 9.869,
"step": 1026
},
{
"epoch": 3.0,
"step": 1026,
"total_flos": 1.4063349020358083e+18,
"train_loss": 0.1461177020974559,
"train_runtime": 16425.358,
"train_samples_per_second": 15.999,
"train_steps_per_second": 0.062
}
],
"max_steps": 1026,
"num_train_epochs": 3,
"total_flos": 1.4063349020358083e+18,
"trial_name": null,
"trial_params": null
}