{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 15.0,
"eval_steps": 500,
"global_step": 39375,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.19,
"learning_rate": 0.0009873015873015875,
"loss": 0.4962,
"step": 500
},
{
"epoch": 0.38,
"learning_rate": 0.0009746031746031746,
"loss": 0.5629,
"step": 1000
},
{
"epoch": 0.57,
"learning_rate": 0.0009619047619047619,
"loss": 0.6018,
"step": 1500
},
{
"epoch": 0.76,
"learning_rate": 0.0009492063492063491,
"loss": 0.6211,
"step": 2000
},
{
"epoch": 0.95,
"learning_rate": 0.0009365079365079366,
"loss": 0.6306,
"step": 2500
},
{
"epoch": 1.0,
"eval_accuracy": 0.6565389538864577,
"eval_loss": 1.9168750047683716,
"eval_runtime": 5346.8612,
"eval_samples_per_second": 31.412,
"eval_steps_per_second": 0.123,
"step": 2625
},
{
"epoch": 1.14,
"learning_rate": 0.0009238095238095239,
"loss": 0.5492,
"step": 3000
},
{
"epoch": 1.33,
"learning_rate": 0.0009111111111111111,
"loss": 0.5403,
"step": 3500
},
{
"epoch": 1.52,
"learning_rate": 0.0008984126984126985,
"loss": 0.5678,
"step": 4000
},
{
"epoch": 1.71,
"learning_rate": 0.0008857142857142857,
"loss": 0.58,
"step": 4500
},
{
"epoch": 1.9,
"learning_rate": 0.000873015873015873,
"loss": 0.5977,
"step": 5000
},
{
"epoch": 2.0,
"eval_accuracy": 0.6613973981125897,
"eval_loss": 1.911996841430664,
"eval_runtime": 2172.477,
"eval_samples_per_second": 77.31,
"eval_steps_per_second": 0.302,
"step": 5250
},
{
"epoch": 2.1,
"learning_rate": 0.0008603174603174603,
"loss": 0.5428,
"step": 5500
},
{
"epoch": 2.29,
"learning_rate": 0.0008476190476190476,
"loss": 0.5113,
"step": 6000
},
{
"epoch": 2.48,
"learning_rate": 0.000834920634920635,
"loss": 0.5199,
"step": 6500
},
{
"epoch": 2.67,
"learning_rate": 0.0008222222222222222,
"loss": 0.5418,
"step": 7000
},
{
"epoch": 2.86,
"learning_rate": 0.0008095238095238096,
"loss": 0.5472,
"step": 7500
},
{
"epoch": 3.0,
"eval_accuracy": 0.6635348754130571,
"eval_loss": 1.9409977197647095,
"eval_runtime": 2183.1517,
"eval_samples_per_second": 76.932,
"eval_steps_per_second": 0.301,
"step": 7875
},
{
"epoch": 3.05,
"learning_rate": 0.0007968253968253968,
"loss": 0.5319,
"step": 8000
},
{
"epoch": 3.24,
"learning_rate": 0.0007841269841269841,
"loss": 0.4719,
"step": 8500
},
{
"epoch": 3.43,
"learning_rate": 0.0007714285714285715,
"loss": 0.4803,
"step": 9000
},
{
"epoch": 3.62,
"learning_rate": 0.0007587301587301587,
"loss": 0.4889,
"step": 9500
},
{
"epoch": 3.81,
"learning_rate": 0.000746031746031746,
"loss": 0.5079,
"step": 10000
},
{
"epoch": 4.0,
"learning_rate": 0.0007333333333333333,
"loss": 0.513,
"step": 10500
},
{
"epoch": 4.0,
"eval_accuracy": 0.6678634157958977,
"eval_loss": 1.9818364381790161,
"eval_runtime": 2205.4104,
"eval_samples_per_second": 76.156,
"eval_steps_per_second": 0.298,
"step": 10500
},
{
"epoch": 4.19,
"learning_rate": 0.0007206349206349207,
"loss": 0.4391,
"step": 11000
},
{
"epoch": 4.38,
"learning_rate": 0.000707936507936508,
"loss": 0.452,
"step": 11500
},
{
"epoch": 4.57,
"learning_rate": 0.0006952380952380952,
"loss": 0.4703,
"step": 12000
},
{
"epoch": 4.76,
"learning_rate": 0.0006825396825396826,
"loss": 0.4698,
"step": 12500
},
{
"epoch": 4.95,
"learning_rate": 0.0006698412698412698,
"loss": 0.4714,
"step": 13000
},
{
"epoch": 5.0,
"eval_accuracy": 0.6711261945163883,
"eval_loss": 1.9427947998046875,
"eval_runtime": 5170.4,
"eval_samples_per_second": 32.484,
"eval_steps_per_second": 0.127,
"step": 13125
},
{
"epoch": 5.14,
"learning_rate": 0.0006571428571428571,
"loss": 0.4337,
"step": 13500
},
{
"epoch": 5.33,
"learning_rate": 0.0006444444444444444,
"loss": 0.4168,
"step": 14000
},
{
"epoch": 5.52,
"learning_rate": 0.0006317460317460318,
"loss": 0.4274,
"step": 14500
},
{
"epoch": 5.71,
"learning_rate": 0.0006190476190476191,
"loss": 0.437,
"step": 15000
},
{
"epoch": 5.9,
"learning_rate": 0.0006063492063492063,
"loss": 0.4378,
"step": 15500
},
{
"epoch": 6.0,
"eval_accuracy": 0.6727397219493316,
"eval_loss": 2.0051467418670654,
"eval_runtime": 2183.8633,
"eval_samples_per_second": 76.907,
"eval_steps_per_second": 0.301,
"step": 15750
},
{
"epoch": 6.1,
"learning_rate": 0.0005936507936507937,
"loss": 0.4139,
"step": 16000
},
{
"epoch": 6.29,
"learning_rate": 0.000580952380952381,
"loss": 0.3956,
"step": 16500
},
{
"epoch": 6.48,
"learning_rate": 0.0005682539682539682,
"loss": 0.3989,
"step": 17000
},
{
"epoch": 6.67,
"learning_rate": 0.0005555555555555556,
"loss": 0.405,
"step": 17500
},
{
"epoch": 6.86,
"learning_rate": 0.0005428571428571428,
"loss": 0.4105,
"step": 18000
},
{
"epoch": 7.0,
"eval_accuracy": 0.677752969545414,
"eval_loss": 1.9916749000549316,
"eval_runtime": 2143.2651,
"eval_samples_per_second": 78.364,
"eval_steps_per_second": 0.307,
"step": 18375
},
{
"epoch": 7.05,
"learning_rate": 0.0005301587301587302,
"loss": 0.405,
"step": 18500
},
{
"epoch": 7.24,
"learning_rate": 0.0005174603174603176,
"loss": 0.3653,
"step": 19000
},
{
"epoch": 7.43,
"learning_rate": 0.0005047619047619048,
"loss": 0.3694,
"step": 19500
},
{
"epoch": 7.62,
"learning_rate": 0.000492063492063492,
"loss": 0.3766,
"step": 20000
},
{
"epoch": 7.81,
"learning_rate": 0.0004793650793650794,
"loss": 0.3774,
"step": 20500
},
{
"epoch": 8.0,
"learning_rate": 0.00046666666666666666,
"loss": 0.3837,
"step": 21000
},
{
"epoch": 8.0,
"eval_accuracy": 0.679211693608407,
"eval_loss": 2.041311025619507,
"eval_runtime": 2155.99,
"eval_samples_per_second": 77.902,
"eval_steps_per_second": 0.305,
"step": 21000
},
{
"epoch": 8.19,
"learning_rate": 0.000453968253968254,
"loss": 0.3406,
"step": 21500
},
{
"epoch": 8.38,
"learning_rate": 0.00044126984126984127,
"loss": 0.3469,
"step": 22000
},
{
"epoch": 8.57,
"learning_rate": 0.00042857142857142855,
"loss": 0.3523,
"step": 22500
},
{
"epoch": 8.76,
"learning_rate": 0.0004158730158730159,
"loss": 0.3559,
"step": 23000
},
{
"epoch": 8.95,
"learning_rate": 0.0004031746031746032,
"loss": 0.3602,
"step": 23500
},
{
"epoch": 9.0,
"eval_accuracy": 0.6870054478878271,
"eval_loss": 2.0737974643707275,
"eval_runtime": 2167.8038,
"eval_samples_per_second": 77.477,
"eval_steps_per_second": 0.303,
"step": 23625
},
{
"epoch": 9.14,
"learning_rate": 0.0003904761904761905,
"loss": 0.3172,
"step": 24000
},
{
"epoch": 9.33,
"learning_rate": 0.00037777777777777777,
"loss": 0.325,
"step": 24500
},
{
"epoch": 9.52,
"learning_rate": 0.00036507936507936505,
"loss": 0.3255,
"step": 25000
},
{
"epoch": 9.71,
"learning_rate": 0.00035238095238095243,
"loss": 0.3277,
"step": 25500
},
{
"epoch": 9.9,
"learning_rate": 0.0003396825396825397,
"loss": 0.336,
"step": 26000
},
{
"epoch": 10.0,
"eval_accuracy": 0.6875770295614897,
"eval_loss": 2.087229013442993,
"eval_runtime": 4594.6085,
"eval_samples_per_second": 36.555,
"eval_steps_per_second": 0.143,
"step": 26250
},
{
"epoch": 10.1,
"learning_rate": 0.000326984126984127,
"loss": 0.3184,
"step": 26500
},
{
"epoch": 10.29,
"learning_rate": 0.00031428571428571427,
"loss": 0.3072,
"step": 27000
},
{
"epoch": 10.48,
"learning_rate": 0.0003015873015873016,
"loss": 0.3024,
"step": 27500
},
{
"epoch": 10.67,
"learning_rate": 0.0002888888888888889,
"loss": 0.3112,
"step": 28000
},
{
"epoch": 10.86,
"learning_rate": 0.0002761904761904762,
"loss": 0.3057,
"step": 28500
},
{
"epoch": 11.0,
"eval_accuracy": 0.6893870381947546,
"eval_loss": 2.1162524223327637,
"eval_runtime": 1647.7962,
"eval_samples_per_second": 101.927,
"eval_steps_per_second": 0.399,
"step": 28875
},
{
"epoch": 11.05,
"learning_rate": 0.0002634920634920635,
"loss": 0.2991,
"step": 29000
},
{
"epoch": 11.24,
"learning_rate": 0.0002507936507936508,
"loss": 0.2861,
"step": 29500
},
{
"epoch": 11.43,
"learning_rate": 0.0002380952380952381,
"loss": 0.2849,
"step": 30000
},
{
"epoch": 11.62,
"learning_rate": 0.0002253968253968254,
"loss": 0.2897,
"step": 30500
},
{
"epoch": 11.81,
"learning_rate": 0.0002126984126984127,
"loss": 0.2864,
"step": 31000
},
{
"epoch": 12.0,
"learning_rate": 0.0002,
"loss": 0.2856,
"step": 31500
},
{
"epoch": 12.0,
"eval_accuracy": 0.6936441308683874,
"eval_loss": 2.1104633808135986,
"eval_runtime": 1643.6927,
"eval_samples_per_second": 102.182,
"eval_steps_per_second": 0.4,
"step": 31500
},
{
"epoch": 12.19,
"learning_rate": 0.00018730158730158731,
"loss": 0.2704,
"step": 32000
},
{
"epoch": 12.38,
"learning_rate": 0.0001746031746031746,
"loss": 0.2688,
"step": 32500
},
{
"epoch": 12.57,
"learning_rate": 0.00016190476190476192,
"loss": 0.2634,
"step": 33000
},
{
"epoch": 12.76,
"learning_rate": 0.0001492063492063492,
"loss": 0.2713,
"step": 33500
},
{
"epoch": 12.95,
"learning_rate": 0.0001365079365079365,
"loss": 0.2704,
"step": 34000
},
{
"epoch": 13.0,
"eval_accuracy": 0.6964782233336311,
"eval_loss": 2.168531894683838,
"eval_runtime": 1642.8004,
"eval_samples_per_second": 102.237,
"eval_steps_per_second": 0.4,
"step": 34125
},
{
"epoch": 13.14,
"learning_rate": 0.0001238095238095238,
"loss": 0.2586,
"step": 34500
},
{
"epoch": 13.33,
"learning_rate": 0.0001111111111111111,
"loss": 0.2544,
"step": 35000
},
{
"epoch": 13.52,
"learning_rate": 9.841269841269841e-05,
"loss": 0.2483,
"step": 35500
},
{
"epoch": 13.71,
"learning_rate": 8.571428571428571e-05,
"loss": 0.2612,
"step": 36000
},
{
"epoch": 13.9,
"learning_rate": 7.301587301587302e-05,
"loss": 0.2503,
"step": 36500
},
{
"epoch": 14.0,
"eval_accuracy": 0.699431395314221,
"eval_loss": 2.162724733352661,
"eval_runtime": 1648.5219,
"eval_samples_per_second": 101.882,
"eval_steps_per_second": 0.399,
"step": 36750
},
{
"epoch": 14.1,
"learning_rate": 6.031746031746032e-05,
"loss": 0.2471,
"step": 37000
},
{
"epoch": 14.29,
"learning_rate": 4.761904761904762e-05,
"loss": 0.2452,
"step": 37500
},
{
"epoch": 14.48,
"learning_rate": 3.4920634920634925e-05,
"loss": 0.2433,
"step": 38000
},
{
"epoch": 14.67,
"learning_rate": 2.2222222222222223e-05,
"loss": 0.2392,
"step": 38500
},
{
"epoch": 14.86,
"learning_rate": 9.523809523809525e-06,
"loss": 0.2362,
"step": 39000
},
{
"epoch": 15.0,
"eval_accuracy": 0.7000208389151856,
"eval_loss": 2.1793291568756104,
"eval_runtime": 1654.6492,
"eval_samples_per_second": 101.505,
"eval_steps_per_second": 0.397,
"step": 39375
},
{
"epoch": 15.0,
"step": 39375,
"total_flos": 6.872831806674565e+20,
"train_loss": 0.1121770017593626,
"train_runtime": 63366.2415,
"train_samples_per_second": 159.032,
"train_steps_per_second": 0.621
}
],
"logging_steps": 500,
"max_steps": 39375,
"num_train_epochs": 15,
"save_steps": 500,
"total_flos": 6.872831806674565e+20,
"trial_name": null,
"trial_params": null
}