mt5_base-qg-ap-peft-oficial / trainer_state.json
tiagoblima's picture
End of training
49c1c0b verified
{
"best_metric": 6.692422866821289,
"best_model_checkpoint": "/temp/mt5_base-qg-ap-peft-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": 14.059,
"step": 100
},
{
"epoch": 0.08,
"learning_rate": 9.831009716941276e-05,
"loss": 13.141,
"step": 200
},
{
"epoch": 0.13,
"learning_rate": 9.746514575411914e-05,
"loss": 12.4595,
"step": 300
},
{
"epoch": 0.17,
"learning_rate": 9.662019433882552e-05,
"loss": 11.9881,
"step": 400
},
{
"epoch": 0.21,
"learning_rate": 9.57752429235319e-05,
"loss": 11.6757,
"step": 500
},
{
"epoch": 0.25,
"learning_rate": 9.493029150823828e-05,
"loss": 11.397,
"step": 600
},
{
"epoch": 0.3,
"learning_rate": 9.408534009294467e-05,
"loss": 11.3676,
"step": 700
},
{
"epoch": 0.34,
"learning_rate": 9.324038867765104e-05,
"loss": 11.213,
"step": 800
},
{
"epoch": 0.38,
"learning_rate": 9.239543726235742e-05,
"loss": 11.1506,
"step": 900
},
{
"epoch": 0.42,
"learning_rate": 9.15504858470638e-05,
"loss": 11.0655,
"step": 1000
},
{
"epoch": 0.46,
"learning_rate": 9.070553443177018e-05,
"loss": 10.8905,
"step": 1100
},
{
"epoch": 0.51,
"learning_rate": 8.986058301647655e-05,
"loss": 10.8484,
"step": 1200
},
{
"epoch": 0.55,
"learning_rate": 8.901563160118294e-05,
"loss": 10.6917,
"step": 1300
},
{
"epoch": 0.59,
"learning_rate": 8.817068018588932e-05,
"loss": 10.7003,
"step": 1400
},
{
"epoch": 0.63,
"learning_rate": 8.732572877059569e-05,
"loss": 10.6591,
"step": 1500
},
{
"epoch": 0.68,
"learning_rate": 8.648077735530208e-05,
"loss": 10.6206,
"step": 1600
},
{
"epoch": 0.72,
"learning_rate": 8.563582594000845e-05,
"loss": 10.5158,
"step": 1700
},
{
"epoch": 0.76,
"learning_rate": 8.479087452471483e-05,
"loss": 10.5008,
"step": 1800
},
{
"epoch": 0.8,
"learning_rate": 8.394592310942122e-05,
"loss": 10.5215,
"step": 1900
},
{
"epoch": 0.84,
"learning_rate": 8.31009716941276e-05,
"loss": 10.3384,
"step": 2000
},
{
"epoch": 0.89,
"learning_rate": 8.225602027883397e-05,
"loss": 10.3644,
"step": 2100
},
{
"epoch": 0.93,
"learning_rate": 8.141106886354036e-05,
"loss": 10.2678,
"step": 2200
},
{
"epoch": 0.97,
"learning_rate": 8.056611744824673e-05,
"loss": 10.2785,
"step": 2300
},
{
"epoch": 1.0,
"eval_loss": 7.456560134887695,
"eval_runtime": 60.5566,
"eval_samples_per_second": 179.848,
"eval_steps_per_second": 2.824,
"step": 2367
},
{
"epoch": 1.01,
"learning_rate": 7.972116603295311e-05,
"loss": 10.1938,
"step": 2400
},
{
"epoch": 1.06,
"learning_rate": 7.887621461765948e-05,
"loss": 10.2523,
"step": 2500
},
{
"epoch": 1.1,
"learning_rate": 7.803126320236587e-05,
"loss": 10.2002,
"step": 2600
},
{
"epoch": 1.14,
"learning_rate": 7.718631178707225e-05,
"loss": 10.1346,
"step": 2700
},
{
"epoch": 1.18,
"learning_rate": 7.634136037177862e-05,
"loss": 10.0902,
"step": 2800
},
{
"epoch": 1.23,
"learning_rate": 7.5496408956485e-05,
"loss": 10.0808,
"step": 2900
},
{
"epoch": 1.27,
"learning_rate": 7.465145754119138e-05,
"loss": 10.1033,
"step": 3000
},
{
"epoch": 1.31,
"learning_rate": 7.380650612589776e-05,
"loss": 10.0139,
"step": 3100
},
{
"epoch": 1.35,
"learning_rate": 7.296155471060415e-05,
"loss": 10.0463,
"step": 3200
},
{
"epoch": 1.39,
"learning_rate": 7.211660329531053e-05,
"loss": 10.0505,
"step": 3300
},
{
"epoch": 1.44,
"learning_rate": 7.12716518800169e-05,
"loss": 9.9703,
"step": 3400
},
{
"epoch": 1.48,
"learning_rate": 7.042670046472329e-05,
"loss": 9.9231,
"step": 3500
},
{
"epoch": 1.52,
"learning_rate": 6.958174904942966e-05,
"loss": 9.8841,
"step": 3600
},
{
"epoch": 1.56,
"learning_rate": 6.873679763413604e-05,
"loss": 9.9664,
"step": 3700
},
{
"epoch": 1.61,
"learning_rate": 6.789184621884243e-05,
"loss": 9.8227,
"step": 3800
},
{
"epoch": 1.65,
"learning_rate": 6.70468948035488e-05,
"loss": 9.8745,
"step": 3900
},
{
"epoch": 1.69,
"learning_rate": 6.620194338825518e-05,
"loss": 9.7746,
"step": 4000
},
{
"epoch": 1.73,
"learning_rate": 6.535699197296155e-05,
"loss": 9.8495,
"step": 4100
},
{
"epoch": 1.77,
"learning_rate": 6.451204055766794e-05,
"loss": 9.8663,
"step": 4200
},
{
"epoch": 1.82,
"learning_rate": 6.366708914237431e-05,
"loss": 9.7799,
"step": 4300
},
{
"epoch": 1.86,
"learning_rate": 6.282213772708069e-05,
"loss": 9.7143,
"step": 4400
},
{
"epoch": 1.9,
"learning_rate": 6.197718631178708e-05,
"loss": 9.8438,
"step": 4500
},
{
"epoch": 1.94,
"learning_rate": 6.113223489649346e-05,
"loss": 9.6942,
"step": 4600
},
{
"epoch": 1.99,
"learning_rate": 6.028728348119983e-05,
"loss": 9.7078,
"step": 4700
},
{
"epoch": 2.0,
"eval_loss": 7.058051586151123,
"eval_runtime": 60.5565,
"eval_samples_per_second": 179.849,
"eval_steps_per_second": 2.824,
"step": 4734
},
{
"epoch": 2.03,
"learning_rate": 5.944233206590622e-05,
"loss": 9.7346,
"step": 4800
},
{
"epoch": 2.07,
"learning_rate": 5.859738065061259e-05,
"loss": 9.6838,
"step": 4900
},
{
"epoch": 2.11,
"learning_rate": 5.775242923531897e-05,
"loss": 9.6787,
"step": 5000
},
{
"epoch": 2.15,
"learning_rate": 5.6907477820025357e-05,
"loss": 9.67,
"step": 5100
},
{
"epoch": 2.2,
"learning_rate": 5.606252640473173e-05,
"loss": 9.665,
"step": 5200
},
{
"epoch": 2.24,
"learning_rate": 5.521757498943811e-05,
"loss": 9.7004,
"step": 5300
},
{
"epoch": 2.28,
"learning_rate": 5.437262357414449e-05,
"loss": 9.6461,
"step": 5400
},
{
"epoch": 2.32,
"learning_rate": 5.3527672158850874e-05,
"loss": 9.5778,
"step": 5500
},
{
"epoch": 2.37,
"learning_rate": 5.2682720743557245e-05,
"loss": 9.593,
"step": 5600
},
{
"epoch": 2.41,
"learning_rate": 5.183776932826363e-05,
"loss": 9.6117,
"step": 5700
},
{
"epoch": 2.45,
"learning_rate": 5.099281791297e-05,
"loss": 9.5917,
"step": 5800
},
{
"epoch": 2.49,
"learning_rate": 5.0147866497676385e-05,
"loss": 9.4874,
"step": 5900
},
{
"epoch": 2.53,
"learning_rate": 4.930291508238277e-05,
"loss": 9.6035,
"step": 6000
},
{
"epoch": 2.58,
"learning_rate": 4.845796366708915e-05,
"loss": 9.5254,
"step": 6100
},
{
"epoch": 2.62,
"learning_rate": 4.7613012251795525e-05,
"loss": 9.4565,
"step": 6200
},
{
"epoch": 2.66,
"learning_rate": 4.67680608365019e-05,
"loss": 9.5248,
"step": 6300
},
{
"epoch": 2.7,
"learning_rate": 4.592310942120828e-05,
"loss": 9.4883,
"step": 6400
},
{
"epoch": 2.75,
"learning_rate": 4.507815800591466e-05,
"loss": 9.5128,
"step": 6500
},
{
"epoch": 2.79,
"learning_rate": 4.423320659062104e-05,
"loss": 9.508,
"step": 6600
},
{
"epoch": 2.83,
"learning_rate": 4.338825517532742e-05,
"loss": 9.4925,
"step": 6700
},
{
"epoch": 2.87,
"learning_rate": 4.25433037600338e-05,
"loss": 9.42,
"step": 6800
},
{
"epoch": 2.92,
"learning_rate": 4.1698352344740176e-05,
"loss": 9.4666,
"step": 6900
},
{
"epoch": 2.96,
"learning_rate": 4.085340092944656e-05,
"loss": 9.4949,
"step": 7000
},
{
"epoch": 3.0,
"learning_rate": 4.000844951415294e-05,
"loss": 9.4851,
"step": 7100
},
{
"epoch": 3.0,
"eval_loss": 6.835981369018555,
"eval_runtime": 60.5893,
"eval_samples_per_second": 179.751,
"eval_steps_per_second": 2.822,
"step": 7101
},
{
"epoch": 3.04,
"learning_rate": 3.916349809885932e-05,
"loss": 9.4286,
"step": 7200
},
{
"epoch": 3.08,
"learning_rate": 3.83185466835657e-05,
"loss": 9.4941,
"step": 7300
},
{
"epoch": 3.13,
"learning_rate": 3.747359526827208e-05,
"loss": 9.435,
"step": 7400
},
{
"epoch": 3.17,
"learning_rate": 3.6628643852978456e-05,
"loss": 9.2994,
"step": 7500
},
{
"epoch": 3.21,
"learning_rate": 3.5783692437684833e-05,
"loss": 9.4189,
"step": 7600
},
{
"epoch": 3.25,
"learning_rate": 3.493874102239121e-05,
"loss": 9.4034,
"step": 7700
},
{
"epoch": 3.3,
"learning_rate": 3.4093789607097596e-05,
"loss": 9.408,
"step": 7800
},
{
"epoch": 3.34,
"learning_rate": 3.324883819180397e-05,
"loss": 9.3133,
"step": 7900
},
{
"epoch": 3.38,
"learning_rate": 3.240388677651035e-05,
"loss": 9.3927,
"step": 8000
},
{
"epoch": 3.42,
"learning_rate": 3.155893536121673e-05,
"loss": 9.3392,
"step": 8100
},
{
"epoch": 3.46,
"learning_rate": 3.0713983945923107e-05,
"loss": 9.3861,
"step": 8200
},
{
"epoch": 3.51,
"learning_rate": 2.986903253062949e-05,
"loss": 9.3197,
"step": 8300
},
{
"epoch": 3.55,
"learning_rate": 2.9024081115335872e-05,
"loss": 9.26,
"step": 8400
},
{
"epoch": 3.59,
"learning_rate": 2.817912970004225e-05,
"loss": 9.3923,
"step": 8500
},
{
"epoch": 3.63,
"learning_rate": 2.733417828474863e-05,
"loss": 9.245,
"step": 8600
},
{
"epoch": 3.68,
"learning_rate": 2.648922686945501e-05,
"loss": 9.3511,
"step": 8700
},
{
"epoch": 3.72,
"learning_rate": 2.5644275454161386e-05,
"loss": 9.3352,
"step": 8800
},
{
"epoch": 3.76,
"learning_rate": 2.4799324038867768e-05,
"loss": 9.3191,
"step": 8900
},
{
"epoch": 3.8,
"learning_rate": 2.3954372623574145e-05,
"loss": 9.4017,
"step": 9000
},
{
"epoch": 3.84,
"learning_rate": 2.3109421208280523e-05,
"loss": 9.2863,
"step": 9100
},
{
"epoch": 3.89,
"learning_rate": 2.2264469792986904e-05,
"loss": 9.2874,
"step": 9200
},
{
"epoch": 3.93,
"learning_rate": 2.1419518377693285e-05,
"loss": 9.3056,
"step": 9300
},
{
"epoch": 3.97,
"learning_rate": 2.0574566962399663e-05,
"loss": 9.2737,
"step": 9400
},
{
"epoch": 4.0,
"eval_loss": 6.7291483879089355,
"eval_runtime": 60.5622,
"eval_samples_per_second": 179.832,
"eval_steps_per_second": 2.824,
"step": 9468
},
{
"epoch": 4.01,
"learning_rate": 1.9729615547106044e-05,
"loss": 9.2448,
"step": 9500
},
{
"epoch": 4.06,
"learning_rate": 1.8884664131812422e-05,
"loss": 9.2816,
"step": 9600
},
{
"epoch": 4.1,
"learning_rate": 1.80397127165188e-05,
"loss": 9.2011,
"step": 9700
},
{
"epoch": 4.14,
"learning_rate": 1.719476130122518e-05,
"loss": 9.2675,
"step": 9800
},
{
"epoch": 4.18,
"learning_rate": 1.634980988593156e-05,
"loss": 9.3373,
"step": 9900
},
{
"epoch": 4.22,
"learning_rate": 1.550485847063794e-05,
"loss": 9.198,
"step": 10000
},
{
"epoch": 4.27,
"learning_rate": 1.4659907055344319e-05,
"loss": 9.3173,
"step": 10100
},
{
"epoch": 4.31,
"learning_rate": 1.3814955640050698e-05,
"loss": 9.343,
"step": 10200
},
{
"epoch": 4.35,
"learning_rate": 1.2970004224757076e-05,
"loss": 9.1825,
"step": 10300
},
{
"epoch": 4.39,
"learning_rate": 1.2125052809463457e-05,
"loss": 9.2288,
"step": 10400
},
{
"epoch": 4.44,
"learning_rate": 1.1280101394169835e-05,
"loss": 9.2879,
"step": 10500
},
{
"epoch": 4.48,
"learning_rate": 1.0435149978876214e-05,
"loss": 9.2281,
"step": 10600
},
{
"epoch": 4.52,
"learning_rate": 9.590198563582595e-06,
"loss": 9.1823,
"step": 10700
},
{
"epoch": 4.56,
"learning_rate": 8.745247148288973e-06,
"loss": 9.2309,
"step": 10800
},
{
"epoch": 4.6,
"learning_rate": 7.900295732995352e-06,
"loss": 9.2587,
"step": 10900
},
{
"epoch": 4.65,
"learning_rate": 7.055344317701733e-06,
"loss": 9.2312,
"step": 11000
},
{
"epoch": 4.69,
"learning_rate": 6.210392902408112e-06,
"loss": 9.1884,
"step": 11100
},
{
"epoch": 4.73,
"learning_rate": 5.3654414871144915e-06,
"loss": 9.2054,
"step": 11200
},
{
"epoch": 4.77,
"learning_rate": 4.52049007182087e-06,
"loss": 9.3236,
"step": 11300
},
{
"epoch": 4.82,
"learning_rate": 3.67553865652725e-06,
"loss": 9.2313,
"step": 11400
},
{
"epoch": 4.86,
"learning_rate": 2.830587241233629e-06,
"loss": 9.1894,
"step": 11500
},
{
"epoch": 4.9,
"learning_rate": 1.9856358259400087e-06,
"loss": 9.3008,
"step": 11600
},
{
"epoch": 4.94,
"learning_rate": 1.140684410646388e-06,
"loss": 9.2425,
"step": 11700
},
{
"epoch": 4.99,
"learning_rate": 2.957329953527672e-07,
"loss": 9.3588,
"step": 11800
},
{
"epoch": 5.0,
"eval_loss": 6.692422866821289,
"eval_runtime": 60.6225,
"eval_samples_per_second": 179.653,
"eval_steps_per_second": 2.821,
"step": 11835
},
{
"epoch": 5.0,
"step": 11835,
"total_flos": 6.784131924335002e+17,
"train_loss": 9.848386054609112,
"train_runtime": 3179.2752,
"train_samples_per_second": 119.087,
"train_steps_per_second": 3.723
}
],
"logging_steps": 100,
"max_steps": 11835,
"num_train_epochs": 5,
"save_steps": 500,
"total_flos": 6.784131924335002e+17,
"trial_name": null,
"trial_params": null
}