Alex-xu's picture
Training in progress, step 23500, checkpoint
c919504
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.806804093988756,
"eval_steps": 500,
"global_step": 23500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"ep_loss": 4.4238,
"epoch": 0.04,
"learning_rate": 2.5e-05,
"loss": 11.1467,
"mlm_loss": 6.7229,
"step": 500
},
{
"ep_loss": 0.4305,
"epoch": 0.08,
"learning_rate": 5e-05,
"loss": 2.9841,
"mlm_loss": 2.5537,
"step": 1000
},
{
"ep_loss": 0.3724,
"epoch": 0.12,
"learning_rate": 7.5e-05,
"loss": 2.2371,
"mlm_loss": 1.8647,
"step": 1500
},
{
"ep_loss": 0.364,
"epoch": 0.15,
"learning_rate": 0.0001,
"loss": 1.8477,
"mlm_loss": 1.4837,
"step": 2000
},
{
"ep_loss": 0.3678,
"epoch": 0.19,
"learning_rate": 0.00012495,
"loss": 1.5215,
"mlm_loss": 1.1538,
"step": 2500
},
{
"ep_loss": 0.3617,
"epoch": 0.23,
"learning_rate": 0.00014995,
"loss": 1.4119,
"mlm_loss": 1.0501,
"step": 3000
},
{
"ep_loss": 0.3336,
"epoch": 0.27,
"learning_rate": 0.0001749,
"loss": 1.3027,
"mlm_loss": 0.9691,
"step": 3500
},
{
"ep_loss": 0.3348,
"epoch": 0.31,
"learning_rate": 0.0001999,
"loss": 1.2441,
"mlm_loss": 0.9093,
"step": 4000
},
{
"ep_loss": 0.3348,
"epoch": 0.35,
"learning_rate": 0.0002249,
"loss": 1.1942,
"mlm_loss": 0.8594,
"step": 4500
},
{
"ep_loss": 0.3331,
"epoch": 0.38,
"learning_rate": 0.0002499,
"loss": 1.1466,
"mlm_loss": 0.8135,
"step": 5000
},
{
"ep_loss": 0.3268,
"epoch": 0.42,
"learning_rate": 0.00027489999999999996,
"loss": 1.1067,
"mlm_loss": 0.7799,
"step": 5500
},
{
"ep_loss": 0.3378,
"epoch": 0.46,
"learning_rate": 0.00029985,
"loss": 1.1007,
"mlm_loss": 0.7629,
"step": 6000
},
{
"ep_loss": 0.3033,
"epoch": 0.5,
"learning_rate": 0.00032485,
"loss": 1.028,
"mlm_loss": 0.7246,
"step": 6500
},
{
"ep_loss": 0.2848,
"epoch": 0.54,
"learning_rate": 0.00034985,
"loss": 0.9833,
"mlm_loss": 0.6985,
"step": 7000
},
{
"ep_loss": 0.316,
"epoch": 0.58,
"learning_rate": 0.00037485000000000003,
"loss": 1.0235,
"mlm_loss": 0.7076,
"step": 7500
},
{
"ep_loss": 0.2885,
"epoch": 0.62,
"learning_rate": 0.00039975,
"loss": 0.9673,
"mlm_loss": 0.6789,
"step": 8000
},
{
"ep_loss": 0.274,
"epoch": 0.65,
"learning_rate": 0.00042475000000000005,
"loss": 0.9266,
"mlm_loss": 0.6525,
"step": 8500
},
{
"ep_loss": 0.3008,
"epoch": 0.69,
"learning_rate": 0.0004497,
"loss": 0.9646,
"mlm_loss": 0.6638,
"step": 9000
},
{
"ep_loss": 0.2843,
"epoch": 0.73,
"learning_rate": 0.00047470000000000005,
"loss": 0.9349,
"mlm_loss": 0.6506,
"step": 9500
},
{
"ep_loss": 0.2693,
"epoch": 0.77,
"learning_rate": 0.0004997,
"loss": 0.8901,
"mlm_loss": 0.6208,
"step": 10000
},
{
"ep_loss": 0.2655,
"epoch": 0.81,
"learning_rate": 0.0004995159140796489,
"loss": 0.8783,
"mlm_loss": 0.6128,
"step": 10500
},
{
"ep_loss": 0.2711,
"epoch": 0.85,
"learning_rate": 0.0004990269285042333,
"loss": 0.8794,
"mlm_loss": 0.6083,
"step": 11000
},
{
"ep_loss": 0.2579,
"epoch": 0.88,
"learning_rate": 0.000498536962997805,
"loss": 0.8489,
"mlm_loss": 0.591,
"step": 11500
},
{
"ep_loss": 0.2607,
"epoch": 0.92,
"learning_rate": 0.0004980469974913766,
"loss": 0.8504,
"mlm_loss": 0.5898,
"step": 12000
},
{
"ep_loss": 0.2585,
"epoch": 0.96,
"learning_rate": 0.0004975570319849483,
"loss": 0.8338,
"mlm_loss": 0.5753,
"step": 12500
},
{
"ep_loss": 0.2584,
"epoch": 1.0,
"learning_rate": 0.0004970670664785199,
"loss": 0.8287,
"mlm_loss": 0.5703,
"step": 13000
},
{
"ep_loss": 0.2511,
"epoch": 1.04,
"learning_rate": 0.0004965771009720916,
"loss": 0.8047,
"mlm_loss": 0.5536,
"step": 13500
},
{
"ep_loss": 0.2476,
"epoch": 1.08,
"learning_rate": 0.0004960871354656633,
"loss": 0.7952,
"mlm_loss": 0.5476,
"step": 14000
},
{
"ep_loss": 0.2482,
"epoch": 1.11,
"learning_rate": 0.0004955971699592349,
"loss": 0.7894,
"mlm_loss": 0.5412,
"step": 14500
},
{
"ep_loss": 0.2424,
"epoch": 1.15,
"learning_rate": 0.0004951072044528065,
"loss": 0.7744,
"mlm_loss": 0.5321,
"step": 15000
},
{
"ep_loss": 0.2462,
"epoch": 1.19,
"learning_rate": 0.0004946182188773911,
"loss": 0.7849,
"mlm_loss": 0.5386,
"step": 15500
},
{
"ep_loss": 0.235,
"epoch": 1.23,
"learning_rate": 0.0004941282533709627,
"loss": 0.7563,
"mlm_loss": 0.5213,
"step": 16000
},
{
"ep_loss": 0.2354,
"epoch": 1.27,
"learning_rate": 0.0004936382878645343,
"loss": 0.7513,
"mlm_loss": 0.5159,
"step": 16500
},
{
"ep_loss": 0.2336,
"epoch": 1.31,
"learning_rate": 0.0004931493022891189,
"loss": 0.7454,
"mlm_loss": 0.5118,
"step": 17000
},
{
"ep_loss": 0.233,
"epoch": 1.35,
"learning_rate": 0.0004926603167137033,
"loss": 0.7403,
"mlm_loss": 0.5073,
"step": 17500
},
{
"ep_loss": 0.2345,
"epoch": 1.38,
"learning_rate": 0.0004921713311382879,
"loss": 0.7399,
"mlm_loss": 0.5054,
"step": 18000
},
{
"ep_loss": 0.2258,
"epoch": 1.42,
"learning_rate": 0.0004916823455628724,
"loss": 0.7183,
"mlm_loss": 0.4925,
"step": 18500
},
{
"ep_loss": 0.235,
"epoch": 1.46,
"learning_rate": 0.0004911962997804954,
"loss": 0.7247,
"mlm_loss": 0.4897,
"step": 19000
},
{
"ep_loss": 0.2446,
"epoch": 1.5,
"learning_rate": 0.0004907082941360928,
"loss": 0.7579,
"mlm_loss": 0.5133,
"step": 19500
},
{
"ep_loss": 0.2295,
"epoch": 1.54,
"learning_rate": 0.0004902183286296644,
"loss": 0.7203,
"mlm_loss": 0.4908,
"step": 20000
},
{
"ep_loss": 0.2332,
"epoch": 1.58,
"learning_rate": 0.0004897283631232362,
"loss": 0.739,
"mlm_loss": 0.5058,
"step": 20500
},
{
"ep_loss": 0.2257,
"epoch": 1.61,
"learning_rate": 0.0004892383976168078,
"loss": 0.7104,
"mlm_loss": 0.4847,
"step": 21000
},
{
"ep_loss": 0.2296,
"epoch": 1.65,
"learning_rate": 0.0004887484321103795,
"loss": 0.7109,
"mlm_loss": 0.4813,
"step": 21500
},
{
"ep_loss": 0.2235,
"epoch": 1.69,
"learning_rate": 0.00048825846660395105,
"loss": 0.6995,
"mlm_loss": 0.476,
"step": 22000
},
{
"ep_loss": 0.2221,
"epoch": 1.73,
"learning_rate": 0.00048776850109752277,
"loss": 0.6947,
"mlm_loss": 0.4726,
"step": 22500
},
{
"ep_loss": 0.2251,
"epoch": 1.77,
"learning_rate": 0.00048727951552210723,
"loss": 0.6991,
"mlm_loss": 0.4741,
"step": 23000
},
{
"ep_loss": 0.2253,
"epoch": 1.81,
"learning_rate": 0.00048678955001567895,
"loss": 0.7069,
"mlm_loss": 0.4816,
"step": 23500
}
],
"logging_steps": 500,
"max_steps": 520240,
"num_train_epochs": 40,
"save_steps": 500,
"total_flos": 2.243820589083001e+19,
"trial_name": null,
"trial_params": null
}