mt5_base-qg-ap-oficial / trainer_state.json
tiagoblima's picture
End of training
e07105b verified
raw
history blame
16 kB
{
"best_metric": 1.3846333026885986,
"best_model_checkpoint": "/temp/mt5_base-qg-ap-oficial/checkpoint-11835",
"epoch": 5.0,
"eval_steps": 500,
"global_step": 11835,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.04,
"learning_rate": 9.915504858470639e-05,
"loss": 6.103,
"step": 100
},
{
"epoch": 0.08,
"learning_rate": 9.831009716941276e-05,
"loss": 2.449,
"step": 200
},
{
"epoch": 0.13,
"learning_rate": 9.746514575411914e-05,
"loss": 2.172,
"step": 300
},
{
"epoch": 0.17,
"learning_rate": 9.662019433882552e-05,
"loss": 2.0448,
"step": 400
},
{
"epoch": 0.21,
"learning_rate": 9.57752429235319e-05,
"loss": 1.9746,
"step": 500
},
{
"epoch": 0.25,
"learning_rate": 9.493029150823828e-05,
"loss": 1.9937,
"step": 600
},
{
"epoch": 0.3,
"learning_rate": 9.408534009294467e-05,
"loss": 1.9024,
"step": 700
},
{
"epoch": 0.34,
"learning_rate": 9.324038867765104e-05,
"loss": 1.8813,
"step": 800
},
{
"epoch": 0.38,
"learning_rate": 9.239543726235742e-05,
"loss": 1.861,
"step": 900
},
{
"epoch": 0.42,
"learning_rate": 9.15504858470638e-05,
"loss": 1.8441,
"step": 1000
},
{
"epoch": 0.46,
"learning_rate": 9.070553443177018e-05,
"loss": 1.7791,
"step": 1100
},
{
"epoch": 0.51,
"learning_rate": 8.986058301647655e-05,
"loss": 1.799,
"step": 1200
},
{
"epoch": 0.55,
"learning_rate": 8.901563160118294e-05,
"loss": 1.7552,
"step": 1300
},
{
"epoch": 0.59,
"learning_rate": 8.817068018588932e-05,
"loss": 1.754,
"step": 1400
},
{
"epoch": 0.63,
"learning_rate": 8.732572877059569e-05,
"loss": 1.7482,
"step": 1500
},
{
"epoch": 0.68,
"learning_rate": 8.648077735530208e-05,
"loss": 1.7425,
"step": 1600
},
{
"epoch": 0.72,
"learning_rate": 8.563582594000845e-05,
"loss": 1.7095,
"step": 1700
},
{
"epoch": 0.76,
"learning_rate": 8.479087452471483e-05,
"loss": 1.7102,
"step": 1800
},
{
"epoch": 0.8,
"learning_rate": 8.394592310942122e-05,
"loss": 1.6842,
"step": 1900
},
{
"epoch": 0.84,
"learning_rate": 8.31009716941276e-05,
"loss": 1.7201,
"step": 2000
},
{
"epoch": 0.89,
"learning_rate": 8.225602027883397e-05,
"loss": 1.6671,
"step": 2100
},
{
"epoch": 0.93,
"learning_rate": 8.141106886354036e-05,
"loss": 1.6492,
"step": 2200
},
{
"epoch": 0.97,
"learning_rate": 8.056611744824673e-05,
"loss": 1.6703,
"step": 2300
},
{
"epoch": 1.0,
"eval_loss": 1.4646919965744019,
"eval_runtime": 55.4142,
"eval_samples_per_second": 196.538,
"eval_steps_per_second": 3.086,
"step": 2367
},
{
"epoch": 1.01,
"learning_rate": 7.972116603295311e-05,
"loss": 1.628,
"step": 2400
},
{
"epoch": 1.06,
"learning_rate": 7.887621461765948e-05,
"loss": 1.6018,
"step": 2500
},
{
"epoch": 1.1,
"learning_rate": 7.803126320236587e-05,
"loss": 1.5866,
"step": 2600
},
{
"epoch": 1.14,
"learning_rate": 7.718631178707225e-05,
"loss": 1.5806,
"step": 2700
},
{
"epoch": 1.18,
"learning_rate": 7.634136037177862e-05,
"loss": 1.5696,
"step": 2800
},
{
"epoch": 1.23,
"learning_rate": 7.5496408956485e-05,
"loss": 1.5574,
"step": 2900
},
{
"epoch": 1.27,
"learning_rate": 7.465145754119138e-05,
"loss": 1.5606,
"step": 3000
},
{
"epoch": 1.31,
"learning_rate": 7.380650612589776e-05,
"loss": 1.5492,
"step": 3100
},
{
"epoch": 1.35,
"learning_rate": 7.296155471060415e-05,
"loss": 1.5233,
"step": 3200
},
{
"epoch": 1.39,
"learning_rate": 7.211660329531053e-05,
"loss": 1.5523,
"step": 3300
},
{
"epoch": 1.44,
"learning_rate": 7.12716518800169e-05,
"loss": 1.5383,
"step": 3400
},
{
"epoch": 1.48,
"learning_rate": 7.042670046472329e-05,
"loss": 1.539,
"step": 3500
},
{
"epoch": 1.52,
"learning_rate": 6.958174904942966e-05,
"loss": 1.5204,
"step": 3600
},
{
"epoch": 1.56,
"learning_rate": 6.873679763413604e-05,
"loss": 1.5319,
"step": 3700
},
{
"epoch": 1.61,
"learning_rate": 6.789184621884243e-05,
"loss": 1.5299,
"step": 3800
},
{
"epoch": 1.65,
"learning_rate": 6.70468948035488e-05,
"loss": 1.5009,
"step": 3900
},
{
"epoch": 1.69,
"learning_rate": 6.620194338825518e-05,
"loss": 1.53,
"step": 4000
},
{
"epoch": 1.73,
"learning_rate": 6.535699197296155e-05,
"loss": 1.5182,
"step": 4100
},
{
"epoch": 1.77,
"learning_rate": 6.451204055766794e-05,
"loss": 1.5236,
"step": 4200
},
{
"epoch": 1.82,
"learning_rate": 6.366708914237431e-05,
"loss": 1.505,
"step": 4300
},
{
"epoch": 1.86,
"learning_rate": 6.282213772708069e-05,
"loss": 1.5032,
"step": 4400
},
{
"epoch": 1.9,
"learning_rate": 6.197718631178708e-05,
"loss": 1.5271,
"step": 4500
},
{
"epoch": 1.94,
"learning_rate": 6.113223489649346e-05,
"loss": 1.4871,
"step": 4600
},
{
"epoch": 1.99,
"learning_rate": 6.028728348119983e-05,
"loss": 1.4897,
"step": 4700
},
{
"epoch": 2.0,
"eval_loss": 1.4201267957687378,
"eval_runtime": 55.4329,
"eval_samples_per_second": 196.472,
"eval_steps_per_second": 3.085,
"step": 4734
},
{
"epoch": 2.03,
"learning_rate": 5.944233206590622e-05,
"loss": 1.4431,
"step": 4800
},
{
"epoch": 2.07,
"learning_rate": 5.859738065061259e-05,
"loss": 1.42,
"step": 4900
},
{
"epoch": 2.11,
"learning_rate": 5.775242923531897e-05,
"loss": 1.4434,
"step": 5000
},
{
"epoch": 2.15,
"learning_rate": 5.6907477820025357e-05,
"loss": 1.4304,
"step": 5100
},
{
"epoch": 2.2,
"learning_rate": 5.606252640473173e-05,
"loss": 1.4338,
"step": 5200
},
{
"epoch": 2.24,
"learning_rate": 5.521757498943811e-05,
"loss": 1.4296,
"step": 5300
},
{
"epoch": 2.28,
"learning_rate": 5.437262357414449e-05,
"loss": 1.4361,
"step": 5400
},
{
"epoch": 2.32,
"learning_rate": 5.3527672158850874e-05,
"loss": 1.4272,
"step": 5500
},
{
"epoch": 2.37,
"learning_rate": 5.2682720743557245e-05,
"loss": 1.4215,
"step": 5600
},
{
"epoch": 2.41,
"learning_rate": 5.183776932826363e-05,
"loss": 1.4125,
"step": 5700
},
{
"epoch": 2.45,
"learning_rate": 5.099281791297e-05,
"loss": 1.4064,
"step": 5800
},
{
"epoch": 2.49,
"learning_rate": 5.0147866497676385e-05,
"loss": 1.4074,
"step": 5900
},
{
"epoch": 2.53,
"learning_rate": 4.930291508238277e-05,
"loss": 1.4175,
"step": 6000
},
{
"epoch": 2.58,
"learning_rate": 4.845796366708915e-05,
"loss": 1.4276,
"step": 6100
},
{
"epoch": 2.62,
"learning_rate": 4.7613012251795525e-05,
"loss": 1.4132,
"step": 6200
},
{
"epoch": 2.66,
"learning_rate": 4.67680608365019e-05,
"loss": 1.4019,
"step": 6300
},
{
"epoch": 2.7,
"learning_rate": 4.592310942120828e-05,
"loss": 1.4228,
"step": 6400
},
{
"epoch": 2.75,
"learning_rate": 4.507815800591466e-05,
"loss": 1.4204,
"step": 6500
},
{
"epoch": 2.79,
"learning_rate": 4.423320659062104e-05,
"loss": 1.3967,
"step": 6600
},
{
"epoch": 2.83,
"learning_rate": 4.338825517532742e-05,
"loss": 1.4127,
"step": 6700
},
{
"epoch": 2.87,
"learning_rate": 4.25433037600338e-05,
"loss": 1.4054,
"step": 6800
},
{
"epoch": 2.92,
"learning_rate": 4.1698352344740176e-05,
"loss": 1.3982,
"step": 6900
},
{
"epoch": 2.96,
"learning_rate": 4.085340092944656e-05,
"loss": 1.4325,
"step": 7000
},
{
"epoch": 3.0,
"learning_rate": 4.000844951415294e-05,
"loss": 1.3982,
"step": 7100
},
{
"epoch": 3.0,
"eval_loss": 1.3959996700286865,
"eval_runtime": 55.4627,
"eval_samples_per_second": 196.366,
"eval_steps_per_second": 3.083,
"step": 7101
},
{
"epoch": 3.04,
"learning_rate": 3.916349809885932e-05,
"loss": 1.3626,
"step": 7200
},
{
"epoch": 3.08,
"learning_rate": 3.83185466835657e-05,
"loss": 1.3606,
"step": 7300
},
{
"epoch": 3.13,
"learning_rate": 3.747359526827208e-05,
"loss": 1.341,
"step": 7400
},
{
"epoch": 3.17,
"learning_rate": 3.6628643852978456e-05,
"loss": 1.368,
"step": 7500
},
{
"epoch": 3.21,
"learning_rate": 3.5783692437684833e-05,
"loss": 1.3629,
"step": 7600
},
{
"epoch": 3.25,
"learning_rate": 3.493874102239121e-05,
"loss": 1.3424,
"step": 7700
},
{
"epoch": 3.3,
"learning_rate": 3.4093789607097596e-05,
"loss": 1.3423,
"step": 7800
},
{
"epoch": 3.34,
"learning_rate": 3.324883819180397e-05,
"loss": 1.3487,
"step": 7900
},
{
"epoch": 3.38,
"learning_rate": 3.240388677651035e-05,
"loss": 1.3473,
"step": 8000
},
{
"epoch": 3.42,
"learning_rate": 3.155893536121673e-05,
"loss": 1.3626,
"step": 8100
},
{
"epoch": 3.46,
"learning_rate": 3.0713983945923107e-05,
"loss": 1.3643,
"step": 8200
},
{
"epoch": 3.51,
"learning_rate": 2.986903253062949e-05,
"loss": 1.3493,
"step": 8300
},
{
"epoch": 3.55,
"learning_rate": 2.9024081115335872e-05,
"loss": 1.3346,
"step": 8400
},
{
"epoch": 3.59,
"learning_rate": 2.817912970004225e-05,
"loss": 1.3516,
"step": 8500
},
{
"epoch": 3.63,
"learning_rate": 2.733417828474863e-05,
"loss": 1.3428,
"step": 8600
},
{
"epoch": 3.68,
"learning_rate": 2.648922686945501e-05,
"loss": 1.3295,
"step": 8700
},
{
"epoch": 3.72,
"learning_rate": 2.5644275454161386e-05,
"loss": 1.3409,
"step": 8800
},
{
"epoch": 3.76,
"learning_rate": 2.4799324038867768e-05,
"loss": 1.3548,
"step": 8900
},
{
"epoch": 3.8,
"learning_rate": 2.3954372623574145e-05,
"loss": 1.3524,
"step": 9000
},
{
"epoch": 3.84,
"learning_rate": 2.3109421208280523e-05,
"loss": 1.3256,
"step": 9100
},
{
"epoch": 3.89,
"learning_rate": 2.2264469792986904e-05,
"loss": 1.3438,
"step": 9200
},
{
"epoch": 3.93,
"learning_rate": 2.1419518377693285e-05,
"loss": 1.3524,
"step": 9300
},
{
"epoch": 3.97,
"learning_rate": 2.0574566962399663e-05,
"loss": 1.3304,
"step": 9400
},
{
"epoch": 4.0,
"eval_loss": 1.385326623916626,
"eval_runtime": 55.4341,
"eval_samples_per_second": 196.468,
"eval_steps_per_second": 3.085,
"step": 9468
},
{
"epoch": 4.01,
"learning_rate": 1.9729615547106044e-05,
"loss": 1.3273,
"step": 9500
},
{
"epoch": 4.06,
"learning_rate": 1.8884664131812422e-05,
"loss": 1.2927,
"step": 9600
},
{
"epoch": 4.1,
"learning_rate": 1.80397127165188e-05,
"loss": 1.3069,
"step": 9700
},
{
"epoch": 4.14,
"learning_rate": 1.719476130122518e-05,
"loss": 1.3173,
"step": 9800
},
{
"epoch": 4.18,
"learning_rate": 1.634980988593156e-05,
"loss": 1.3201,
"step": 9900
},
{
"epoch": 4.22,
"learning_rate": 1.550485847063794e-05,
"loss": 1.3126,
"step": 10000
},
{
"epoch": 4.27,
"learning_rate": 1.4659907055344319e-05,
"loss": 1.3047,
"step": 10100
},
{
"epoch": 4.31,
"learning_rate": 1.3814955640050698e-05,
"loss": 1.3137,
"step": 10200
},
{
"epoch": 4.35,
"learning_rate": 1.2970004224757076e-05,
"loss": 1.3029,
"step": 10300
},
{
"epoch": 4.39,
"learning_rate": 1.2125052809463457e-05,
"loss": 1.3095,
"step": 10400
},
{
"epoch": 4.44,
"learning_rate": 1.1280101394169835e-05,
"loss": 1.3075,
"step": 10500
},
{
"epoch": 4.48,
"learning_rate": 1.0435149978876214e-05,
"loss": 1.3002,
"step": 10600
},
{
"epoch": 4.52,
"learning_rate": 9.590198563582595e-06,
"loss": 1.3297,
"step": 10700
},
{
"epoch": 4.56,
"learning_rate": 8.745247148288973e-06,
"loss": 1.3083,
"step": 10800
},
{
"epoch": 4.6,
"learning_rate": 7.900295732995352e-06,
"loss": 1.3077,
"step": 10900
},
{
"epoch": 4.65,
"learning_rate": 7.055344317701733e-06,
"loss": 1.3037,
"step": 11000
},
{
"epoch": 4.69,
"learning_rate": 6.210392902408112e-06,
"loss": 1.3199,
"step": 11100
},
{
"epoch": 4.73,
"learning_rate": 5.3654414871144915e-06,
"loss": 1.2947,
"step": 11200
},
{
"epoch": 4.77,
"learning_rate": 4.52049007182087e-06,
"loss": 1.3035,
"step": 11300
},
{
"epoch": 4.82,
"learning_rate": 3.67553865652725e-06,
"loss": 1.3147,
"step": 11400
},
{
"epoch": 4.86,
"learning_rate": 2.830587241233629e-06,
"loss": 1.304,
"step": 11500
},
{
"epoch": 4.9,
"learning_rate": 1.9856358259400087e-06,
"loss": 1.307,
"step": 11600
},
{
"epoch": 4.94,
"learning_rate": 1.140684410646388e-06,
"loss": 1.2814,
"step": 11700
},
{
"epoch": 4.99,
"learning_rate": 2.957329953527672e-07,
"loss": 1.2909,
"step": 11800
},
{
"epoch": 5.0,
"eval_loss": 1.3846333026885986,
"eval_runtime": 55.4269,
"eval_samples_per_second": 196.493,
"eval_steps_per_second": 3.085,
"step": 11835
},
{
"epoch": 5.0,
"step": 11835,
"total_flos": 4.539708559274803e+17,
"train_loss": 1.524873844560914,
"train_runtime": 7822.2573,
"train_samples_per_second": 48.402,
"train_steps_per_second": 1.513
}
],
"logging_steps": 100,
"max_steps": 11835,
"num_train_epochs": 5,
"save_steps": 500,
"total_flos": 4.539708559274803e+17,
"trial_name": null,
"trial_params": null
}