ko-TextNumbarT / trainer_state.json
lIlBrother's picture
Init: 모델 최초 commit
50eafd8
raw
history blame
39.8 kB
{
"best_metric": 0.040760207921266556,
"best_model_checkpoint": "/data2/bart/temp_workspace/nlp/output_dir/checkpoint-250000",
"epoch": 5.131554394476582,
"global_step": 330000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.03,
"learning_rate": 2.844285714285714e-07,
"loss": 2.9266,
"step": 2000
},
{
"epoch": 0.06,
"learning_rate": 5.701428571428572e-07,
"loss": 0.9726,
"step": 4000
},
{
"epoch": 0.09,
"learning_rate": 8.558571428571428e-07,
"loss": 0.5831,
"step": 6000
},
{
"epoch": 0.12,
"learning_rate": 1.1415714285714287e-06,
"loss": 0.4407,
"step": 8000
},
{
"epoch": 0.16,
"learning_rate": 1.4272857142857143e-06,
"loss": 0.361,
"step": 10000
},
{
"epoch": 0.16,
"eval_bleu": 0.7861044602086867,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0976053345100065,
"eval_loss": 0.24034307897090912,
"eval_reference_length": 688948,
"eval_rouge1": 0.8478815702507803,
"eval_rouge2": 0.770393954032274,
"eval_rougeL": 0.8471496607249986,
"eval_rougeLsum": 0.8471595113950809,
"eval_runtime": 357.1063,
"eval_samples_per_second": 120.051,
"eval_steps_per_second": 10.005,
"eval_translation_length": 756193,
"step": 10000
},
{
"epoch": 0.19,
"learning_rate": 1.712857142857143e-06,
"loss": 0.2936,
"step": 12000
},
{
"epoch": 0.22,
"learning_rate": 1.9985714285714287e-06,
"loss": 0.255,
"step": 14000
},
{
"epoch": 0.25,
"learning_rate": 2.2841428571428574e-06,
"loss": 0.2187,
"step": 16000
},
{
"epoch": 0.28,
"learning_rate": 2.569714285714286e-06,
"loss": 0.1976,
"step": 18000
},
{
"epoch": 0.31,
"learning_rate": 2.8552857142857144e-06,
"loss": 0.1795,
"step": 20000
},
{
"epoch": 0.31,
"eval_bleu": 0.8725717687820272,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0512491508793116,
"eval_loss": 0.12307040393352509,
"eval_reference_length": 688948,
"eval_rouge1": 0.9065960080443258,
"eval_rouge2": 0.8543292052866391,
"eval_rougeL": 0.9063849761037827,
"eval_rougeLsum": 0.9063953832846721,
"eval_runtime": 357.2954,
"eval_samples_per_second": 119.988,
"eval_steps_per_second": 10.0,
"eval_translation_length": 724256,
"step": 20000
},
{
"epoch": 0.34,
"learning_rate": 3.140857142857143e-06,
"loss": 0.1669,
"step": 22000
},
{
"epoch": 0.37,
"learning_rate": 3.4264285714285715e-06,
"loss": 0.151,
"step": 24000
},
{
"epoch": 0.4,
"learning_rate": 3.7121428571428575e-06,
"loss": 0.1387,
"step": 26000
},
{
"epoch": 0.44,
"learning_rate": 3.997714285714286e-06,
"loss": 0.1354,
"step": 28000
},
{
"epoch": 0.47,
"learning_rate": 4.2834285714285715e-06,
"loss": 0.1244,
"step": 30000
},
{
"epoch": 0.47,
"eval_bleu": 0.8899834257776226,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.050031352148493,
"eval_loss": 0.09066224843263626,
"eval_reference_length": 688948,
"eval_rouge1": 0.9246723733764061,
"eval_rouge2": 0.8828517220657879,
"eval_rougeL": 0.9244225429282074,
"eval_rougeLsum": 0.9244091547844906,
"eval_runtime": 347.7759,
"eval_samples_per_second": 123.272,
"eval_steps_per_second": 10.274,
"eval_translation_length": 723417,
"step": 30000
},
{
"epoch": 0.5,
"learning_rate": 4.569e-06,
"loss": 0.1206,
"step": 32000
},
{
"epoch": 0.53,
"learning_rate": 4.854714285714286e-06,
"loss": 0.1136,
"step": 34000
},
{
"epoch": 0.56,
"learning_rate": 5.140285714285715e-06,
"loss": 0.1128,
"step": 36000
},
{
"epoch": 0.59,
"learning_rate": 5.426e-06,
"loss": 0.1048,
"step": 38000
},
{
"epoch": 0.62,
"learning_rate": 5.7115714285714285e-06,
"loss": 0.1012,
"step": 40000
},
{
"epoch": 0.62,
"eval_bleu": 0.9208253246491475,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0275782787670478,
"eval_loss": 0.07363951206207275,
"eval_reference_length": 688948,
"eval_rouge1": 0.9455229652203183,
"eval_rouge2": 0.9101496822244501,
"eval_rougeL": 0.9454059770773995,
"eval_rougeLsum": 0.9454056293148791,
"eval_runtime": 359.949,
"eval_samples_per_second": 119.103,
"eval_steps_per_second": 9.926,
"eval_translation_length": 707948,
"step": 40000
},
{
"epoch": 0.65,
"learning_rate": 5.997285714285714e-06,
"loss": 0.0962,
"step": 42000
},
{
"epoch": 0.68,
"learning_rate": 6.282571428571429e-06,
"loss": 0.0964,
"step": 44000
},
{
"epoch": 0.72,
"learning_rate": 6.568285714285715e-06,
"loss": 0.095,
"step": 46000
},
{
"epoch": 0.75,
"learning_rate": 6.8538571428571434e-06,
"loss": 0.0913,
"step": 48000
},
{
"epoch": 0.78,
"learning_rate": 7.139428571428572e-06,
"loss": 0.0906,
"step": 50000
},
{
"epoch": 0.78,
"eval_bleu": 0.9072012805362628,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0453851379204235,
"eval_loss": 0.06683389842510223,
"eval_reference_length": 688948,
"eval_rouge1": 0.936913196858357,
"eval_rouge2": 0.9021025087676598,
"eval_rougeL": 0.9368038216303922,
"eval_rougeLsum": 0.936786757645766,
"eval_runtime": 350.5992,
"eval_samples_per_second": 122.279,
"eval_steps_per_second": 10.191,
"eval_translation_length": 720216,
"step": 50000
},
{
"epoch": 0.81,
"learning_rate": 7.425142857142857e-06,
"loss": 0.0881,
"step": 52000
},
{
"epoch": 0.84,
"learning_rate": 7.710714285714287e-06,
"loss": 0.088,
"step": 54000
},
{
"epoch": 0.87,
"learning_rate": 7.996428571428572e-06,
"loss": 0.0852,
"step": 56000
},
{
"epoch": 0.9,
"learning_rate": 8.281857142857143e-06,
"loss": 0.0836,
"step": 58000
},
{
"epoch": 0.93,
"learning_rate": 8.567428571428572e-06,
"loss": 0.0837,
"step": 60000
},
{
"epoch": 0.93,
"eval_bleu": 0.9378929040812776,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0159213757787235,
"eval_loss": 0.06079160049557686,
"eval_reference_length": 688948,
"eval_rouge1": 0.9552455374973553,
"eval_rouge2": 0.9234403473524968,
"eval_rougeL": 0.9552044295582562,
"eval_rougeLsum": 0.9551288128727483,
"eval_runtime": 353.7894,
"eval_samples_per_second": 121.177,
"eval_steps_per_second": 10.099,
"eval_translation_length": 699917,
"step": 60000
},
{
"epoch": 0.96,
"learning_rate": 8.853142857142858e-06,
"loss": 0.0839,
"step": 62000
},
{
"epoch": 1.0,
"learning_rate": 9.138571428571429e-06,
"loss": 0.0815,
"step": 64000
},
{
"epoch": 1.03,
"learning_rate": 9.424285714285715e-06,
"loss": 0.0742,
"step": 66000
},
{
"epoch": 1.06,
"learning_rate": 9.71e-06,
"loss": 0.0683,
"step": 68000
},
{
"epoch": 1.09,
"learning_rate": 9.995714285714286e-06,
"loss": 0.067,
"step": 70000
},
{
"epoch": 1.09,
"eval_bleu": 0.9320243750023054,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0232078473266488,
"eval_loss": 0.058820515871047974,
"eval_reference_length": 688948,
"eval_rouge1": 0.9532786206216067,
"eval_rouge2": 0.9224384838543223,
"eval_rougeL": 0.9531782599798171,
"eval_rougeLsum": 0.9531407362403285,
"eval_runtime": 346.6656,
"eval_samples_per_second": 123.667,
"eval_steps_per_second": 10.307,
"eval_translation_length": 704937,
"step": 70000
},
{
"epoch": 1.12,
"learning_rate": 9.998587848362676e-06,
"loss": 0.0686,
"step": 72000
},
{
"epoch": 1.15,
"learning_rate": 9.994257059669239e-06,
"loss": 0.0718,
"step": 74000
},
{
"epoch": 1.18,
"learning_rate": 9.987009765493164e-06,
"loss": 0.0687,
"step": 76000
},
{
"epoch": 1.21,
"learning_rate": 9.976861810464927e-06,
"loss": 0.0691,
"step": 78000
},
{
"epoch": 1.24,
"learning_rate": 9.963798805467935e-06,
"loss": 0.0726,
"step": 80000
},
{
"epoch": 1.24,
"eval_bleu": 0.931299618108372,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.02620517078212,
"eval_loss": 0.05646243691444397,
"eval_reference_length": 688948,
"eval_rouge1": 0.9506828525475279,
"eval_rouge2": 0.9195555913617979,
"eval_rougeL": 0.9505222105099121,
"eval_rougeLsum": 0.9505025253164432,
"eval_runtime": 358.4748,
"eval_samples_per_second": 119.593,
"eval_steps_per_second": 9.967,
"eval_translation_length": 707002,
"step": 80000
},
{
"epoch": 1.28,
"learning_rate": 9.947845785448258e-06,
"loss": 0.0672,
"step": 82000
},
{
"epoch": 1.31,
"learning_rate": 9.929006251362937e-06,
"loss": 0.0692,
"step": 84000
},
{
"epoch": 1.34,
"learning_rate": 9.907279613454706e-06,
"loss": 0.0682,
"step": 86000
},
{
"epoch": 1.37,
"learning_rate": 9.882700272366437e-06,
"loss": 0.065,
"step": 88000
},
{
"epoch": 1.4,
"learning_rate": 9.855257991301662e-06,
"loss": 0.0648,
"step": 90000
},
{
"epoch": 1.4,
"eval_bleu": 0.9337440022630773,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0237202227163733,
"eval_loss": 0.054729994386434555,
"eval_reference_length": 688948,
"eval_rouge1": 0.9562602670349256,
"eval_rouge2": 0.926727457591136,
"eval_rougeL": 0.9561080754823015,
"eval_rougeLsum": 0.9560774432065933,
"eval_runtime": 347.9396,
"eval_samples_per_second": 123.214,
"eval_steps_per_second": 10.269,
"eval_translation_length": 705290,
"step": 90000
},
{
"epoch": 1.43,
"learning_rate": 9.824996220706527e-06,
"loss": 0.0649,
"step": 92000
},
{
"epoch": 1.46,
"learning_rate": 9.79190235716806e-06,
"loss": 0.067,
"step": 94000
},
{
"epoch": 1.49,
"learning_rate": 9.756028799505886e-06,
"loss": 0.068,
"step": 96000
},
{
"epoch": 1.52,
"learning_rate": 9.717380631513947e-06,
"loss": 0.0609,
"step": 98000
},
{
"epoch": 1.56,
"learning_rate": 9.675980400071303e-06,
"loss": 0.0641,
"step": 100000
},
{
"epoch": 1.56,
"eval_bleu": 0.9437228795566622,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0158429954074908,
"eval_loss": 0.05343015119433403,
"eval_reference_length": 688948,
"eval_rouge1": 0.9622428608075103,
"eval_rouge2": 0.9342448516446199,
"eval_rougeL": 0.9621362437715651,
"eval_rougeLsum": 0.9621409491590999,
"eval_runtime": 354.0159,
"eval_samples_per_second": 121.099,
"eval_steps_per_second": 10.093,
"eval_translation_length": 699863,
"step": 100000
},
{
"epoch": 1.59,
"learning_rate": 9.63182950403888e-06,
"loss": 0.0625,
"step": 102000
},
{
"epoch": 1.62,
"learning_rate": 9.58502194786445e-06,
"loss": 0.0617,
"step": 104000
},
{
"epoch": 1.65,
"learning_rate": 9.535491361586628e-06,
"loss": 0.0637,
"step": 106000
},
{
"epoch": 1.68,
"learning_rate": 9.483312176074826e-06,
"loss": 0.0615,
"step": 108000
},
{
"epoch": 1.71,
"learning_rate": 9.428542910099412e-06,
"loss": 0.061,
"step": 110000
},
{
"epoch": 1.71,
"eval_bleu": 0.9451141494751257,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.016459877958859,
"eval_loss": 0.0512821264564991,
"eval_reference_length": 688948,
"eval_rouge1": 0.96485183292355,
"eval_rouge2": 0.937858740593392,
"eval_rougeL": 0.964753163918546,
"eval_rougeLsum": 0.9647545868693599,
"eval_runtime": 350.3148,
"eval_samples_per_second": 122.379,
"eval_steps_per_second": 10.199,
"eval_translation_length": 700288,
"step": 110000
},
{
"epoch": 1.74,
"learning_rate": 9.371219416280765e-06,
"loss": 0.0647,
"step": 112000
},
{
"epoch": 1.77,
"learning_rate": 9.311287160167118e-06,
"loss": 0.0619,
"step": 114000
},
{
"epoch": 1.8,
"learning_rate": 9.24886908181262e-06,
"loss": 0.0634,
"step": 116000
},
{
"epoch": 1.83,
"learning_rate": 9.18393918535506e-06,
"loss": 0.058,
"step": 118000
},
{
"epoch": 1.87,
"learning_rate": 9.116565986234595e-06,
"loss": 0.0616,
"step": 120000
},
{
"epoch": 1.87,
"eval_bleu": 0.9353441526892242,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0252529944204787,
"eval_loss": 0.04904291033744812,
"eval_reference_length": 688948,
"eval_rouge1": 0.9578546390239036,
"eval_rouge2": 0.9297894756880962,
"eval_rougeL": 0.9577587648786969,
"eval_rougeLsum": 0.957745334663435,
"eval_runtime": 360.6437,
"eval_samples_per_second": 118.874,
"eval_steps_per_second": 9.907,
"eval_translation_length": 706346,
"step": 120000
},
{
"epoch": 1.9,
"learning_rate": 9.04689527434882e-06,
"loss": 0.0588,
"step": 122000
},
{
"epoch": 1.93,
"learning_rate": 8.974795071500655e-06,
"loss": 0.0581,
"step": 124000
},
{
"epoch": 1.96,
"learning_rate": 8.90033821949533e-06,
"loss": 0.0571,
"step": 126000
},
{
"epoch": 1.99,
"learning_rate": 8.823603679804848e-06,
"loss": 0.0575,
"step": 128000
},
{
"epoch": 2.02,
"learning_rate": 8.744676297261531e-06,
"loss": 0.0501,
"step": 130000
},
{
"epoch": 2.02,
"eval_bleu": 0.9391629851344814,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0246375633574667,
"eval_loss": 0.04428655281662941,
"eval_reference_length": 688948,
"eval_rouge1": 0.9620654622397073,
"eval_rouge2": 0.9358278388686297,
"eval_rougeL": 0.9619903188380114,
"eval_rougeLsum": 0.9620056201386562,
"eval_runtime": 350.9058,
"eval_samples_per_second": 122.172,
"eval_steps_per_second": 10.182,
"eval_translation_length": 705922,
"step": 130000
},
{
"epoch": 2.05,
"learning_rate": 8.663523200213197e-06,
"loss": 0.0427,
"step": 132000
},
{
"epoch": 2.08,
"learning_rate": 8.580230708782164e-06,
"loss": 0.0437,
"step": 134000
},
{
"epoch": 2.11,
"learning_rate": 8.494890669233825e-06,
"loss": 0.0425,
"step": 136000
},
{
"epoch": 2.15,
"learning_rate": 8.407555964848785e-06,
"loss": 0.0419,
"step": 138000
},
{
"epoch": 2.18,
"learning_rate": 8.318144933677256e-06,
"loss": 0.0442,
"step": 140000
},
{
"epoch": 2.18,
"eval_bleu": 0.9410527501908049,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0222832492437746,
"eval_loss": 0.04393425211310387,
"eval_reference_length": 688948,
"eval_rouge1": 0.9622283105302971,
"eval_rouge2": 0.9363579229073962,
"eval_rougeL": 0.9621524453412997,
"eval_rougeLsum": 0.9621713486883273,
"eval_runtime": 352.9817,
"eval_samples_per_second": 121.454,
"eval_steps_per_second": 10.122,
"eval_translation_length": 704300,
"step": 140000
},
{
"epoch": 2.21,
"learning_rate": 8.226796199304702e-06,
"loss": 0.0445,
"step": 142000
},
{
"epoch": 2.24,
"learning_rate": 8.13361018519741e-06,
"loss": 0.0445,
"step": 144000
},
{
"epoch": 2.27,
"learning_rate": 8.0386440390423e-06,
"loss": 0.0421,
"step": 146000
},
{
"epoch": 2.3,
"learning_rate": 7.941809254954647e-06,
"loss": 0.0434,
"step": 148000
},
{
"epoch": 2.33,
"learning_rate": 7.843256537104586e-06,
"loss": 0.0418,
"step": 150000
},
{
"epoch": 2.33,
"eval_bleu": 0.9452514516559248,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0194572014143302,
"eval_loss": 0.04334454983472824,
"eval_reference_length": 688948,
"eval_rouge1": 0.9667034100796513,
"eval_rouge2": 0.9415234249395421,
"eval_rougeL": 0.966596537202159,
"eval_rougeLsum": 0.9665995305639754,
"eval_runtime": 350.4381,
"eval_samples_per_second": 122.335,
"eval_steps_per_second": 10.196,
"eval_translation_length": 702353,
"step": 150000
},
{
"epoch": 2.36,
"learning_rate": 7.743043437504057e-06,
"loss": 0.0437,
"step": 152000
},
{
"epoch": 2.39,
"learning_rate": 7.641279775661868e-06,
"loss": 0.0426,
"step": 154000
},
{
"epoch": 2.43,
"learning_rate": 7.538027276115405e-06,
"loss": 0.0412,
"step": 156000
},
{
"epoch": 2.46,
"learning_rate": 7.433190045507044e-06,
"loss": 0.0409,
"step": 158000
},
{
"epoch": 2.49,
"learning_rate": 7.326931900431675e-06,
"loss": 0.0419,
"step": 160000
},
{
"epoch": 2.49,
"eval_bleu": 0.9542889216747796,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.010262022678054,
"eval_loss": 0.04232573136687279,
"eval_reference_length": 688948,
"eval_rouge1": 0.9704801328109232,
"eval_rouge2": 0.9456205175624834,
"eval_rougeL": 0.970368187542237,
"eval_rougeLsum": 0.9703795940864277,
"eval_runtime": 357.9392,
"eval_samples_per_second": 119.772,
"eval_steps_per_second": 9.982,
"eval_translation_length": 696018,
"step": 160000
},
{
"epoch": 2.52,
"learning_rate": 7.219369030269242e-06,
"loss": 0.0421,
"step": 162000
},
{
"epoch": 2.55,
"learning_rate": 7.110456637268422e-06,
"loss": 0.0428,
"step": 164000
},
{
"epoch": 2.58,
"learning_rate": 7.000367166717425e-06,
"loss": 0.0401,
"step": 166000
},
{
"epoch": 2.61,
"learning_rate": 6.889110705881452e-06,
"loss": 0.0417,
"step": 168000
},
{
"epoch": 2.64,
"learning_rate": 6.776752160449367e-06,
"loss": 0.042,
"step": 170000
},
{
"epoch": 2.64,
"eval_bleu": 0.9447481654515599,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0186748491903599,
"eval_loss": 0.04124660789966583,
"eval_reference_length": 688948,
"eval_rouge1": 0.9616790997827227,
"eval_rouge2": 0.9356205732284686,
"eval_rougeL": 0.961519978707468,
"eval_rougeLsum": 0.9615393434549021,
"eval_runtime": 350.4802,
"eval_samples_per_second": 122.321,
"eval_steps_per_second": 10.195,
"eval_translation_length": 701814,
"step": 170000
},
{
"epoch": 2.67,
"learning_rate": 6.663300104771134e-06,
"loss": 0.0386,
"step": 172000
},
{
"epoch": 2.71,
"learning_rate": 6.548876728670889e-06,
"loss": 0.0442,
"step": 174000
},
{
"epoch": 2.74,
"learning_rate": 6.433664609478793e-06,
"loss": 0.0417,
"step": 176000
},
{
"epoch": 2.77,
"learning_rate": 6.3175003841713e-06,
"loss": 0.0418,
"step": 178000
},
{
"epoch": 2.8,
"learning_rate": 6.200566775740774e-06,
"loss": 0.0406,
"step": 180000
},
{
"epoch": 2.8,
"eval_bleu": 0.9487462287220352,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0163989154479003,
"eval_loss": 0.04077766090631485,
"eval_reference_length": 688948,
"eval_rouge1": 0.9665622993166698,
"eval_rouge2": 0.9412909286045468,
"eval_rougeL": 0.9663690505539053,
"eval_rougeLsum": 0.9663768593347282,
"eval_runtime": 354.3716,
"eval_samples_per_second": 120.978,
"eval_steps_per_second": 10.083,
"eval_translation_length": 700246,
"step": 180000
},
{
"epoch": 2.83,
"learning_rate": 6.083109012991928e-06,
"loss": 0.0416,
"step": 182000
},
{
"epoch": 2.86,
"learning_rate": 5.964842802508876e-06,
"loss": 0.0404,
"step": 184000
},
{
"epoch": 2.89,
"learning_rate": 5.846013150999504e-06,
"loss": 0.0379,
"step": 186000
},
{
"epoch": 2.92,
"learning_rate": 5.726749225339994e-06,
"loss": 0.0432,
"step": 188000
},
{
"epoch": 2.95,
"learning_rate": 5.607001354527006e-06,
"loss": 0.0393,
"step": 190000
},
{
"epoch": 2.95,
"eval_bleu": 0.9547919266733801,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0112867734575033,
"eval_loss": 0.04095698148012161,
"eval_reference_length": 688948,
"eval_rouge1": 0.9703758390841425,
"eval_rouge2": 0.9461704701896383,
"eval_rougeL": 0.9702899676023662,
"eval_rougeLsum": 0.9703172456040432,
"eval_runtime": 353.5109,
"eval_samples_per_second": 121.272,
"eval_steps_per_second": 10.107,
"eval_translation_length": 696724,
"step": 190000
},
{
"epoch": 2.99,
"learning_rate": 5.486959140088201e-06,
"loss": 0.0384,
"step": 192000
},
{
"epoch": 3.02,
"learning_rate": 5.366572586878771e-06,
"loss": 0.0354,
"step": 194000
},
{
"epoch": 3.05,
"learning_rate": 5.246032307677414e-06,
"loss": 0.0308,
"step": 196000
},
{
"epoch": 3.08,
"learning_rate": 5.125469288643505e-06,
"loss": 0.033,
"step": 198000
},
{
"epoch": 3.11,
"learning_rate": 5.004651973065896e-06,
"loss": 0.0316,
"step": 200000
},
{
"epoch": 3.11,
"eval_bleu": 0.9516992046244281,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.013291278877361,
"eval_loss": 0.04207807779312134,
"eval_reference_length": 688948,
"eval_rouge1": 0.9687692703791779,
"eval_rouge2": 0.9438025496007139,
"eval_rougeL": 0.9686298083794034,
"eval_rougeLsum": 0.9686054829671312,
"eval_runtime": 358.3539,
"eval_samples_per_second": 119.633,
"eval_steps_per_second": 9.971,
"eval_translation_length": 698105,
"step": 200000
},
{
"epoch": 3.14,
"learning_rate": 4.883831940867018e-06,
"loss": 0.0332,
"step": 202000
},
{
"epoch": 3.17,
"learning_rate": 4.763079747543336e-06,
"loss": 0.03,
"step": 204000
},
{
"epoch": 3.2,
"learning_rate": 4.642526169588442e-06,
"loss": 0.0313,
"step": 206000
},
{
"epoch": 3.23,
"learning_rate": 4.52218113759159e-06,
"loss": 0.0319,
"step": 208000
},
{
"epoch": 3.27,
"learning_rate": 4.402114859405383e-06,
"loss": 0.0313,
"step": 210000
},
{
"epoch": 3.27,
"eval_bleu": 0.952314893533454,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0132433797616076,
"eval_loss": 0.04261790215969086,
"eval_reference_length": 688948,
"eval_rouge1": 0.9687049595684978,
"eval_rouge2": 0.9435984713700241,
"eval_rougeL": 0.9686312046722289,
"eval_rougeLsum": 0.9685702469811763,
"eval_runtime": 353.4407,
"eval_samples_per_second": 121.296,
"eval_steps_per_second": 10.109,
"eval_translation_length": 698072,
"step": 210000
},
{
"epoch": 3.3,
"learning_rate": 4.2823973802607795e-06,
"loss": 0.0322,
"step": 212000
},
{
"epoch": 3.33,
"learning_rate": 4.163038979034976e-06,
"loss": 0.0312,
"step": 214000
},
{
"epoch": 3.36,
"learning_rate": 4.0441693394762706e-06,
"loss": 0.0314,
"step": 216000
},
{
"epoch": 3.39,
"learning_rate": 3.925916882841615e-06,
"loss": 0.0332,
"step": 218000
},
{
"epoch": 3.42,
"learning_rate": 3.8083497076975863e-06,
"loss": 0.0302,
"step": 220000
},
{
"epoch": 3.42,
"eval_bleu": 0.9510751957840793,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.014170880821194,
"eval_loss": 0.04271363466978073,
"eval_reference_length": 688948,
"eval_rouge1": 0.9682283271173224,
"eval_rouge2": 0.9430416842293828,
"eval_rougeL": 0.9681608428660635,
"eval_rougeLsum": 0.9681434321637754,
"eval_runtime": 354.0647,
"eval_samples_per_second": 121.082,
"eval_steps_per_second": 10.091,
"eval_translation_length": 698711,
"step": 220000
},
{
"epoch": 3.45,
"learning_rate": 3.691418722192835e-06,
"loss": 0.0323,
"step": 222000
},
{
"epoch": 3.48,
"learning_rate": 3.5751932368863875e-06,
"loss": 0.0302,
"step": 224000
},
{
"epoch": 3.51,
"learning_rate": 3.4597997986064915e-06,
"loss": 0.0312,
"step": 226000
},
{
"epoch": 3.55,
"learning_rate": 3.3453057938715767e-06,
"loss": 0.0309,
"step": 228000
},
{
"epoch": 3.58,
"learning_rate": 3.2318345952925634e-06,
"loss": 0.0305,
"step": 230000
},
{
"epoch": 3.58,
"eval_bleu": 0.9477212328461616,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.017851855292417,
"eval_loss": 0.042060352861881256,
"eval_reference_length": 688948,
"eval_rouge1": 0.9686642355024724,
"eval_rouge2": 0.9435154406416524,
"eval_rougeL": 0.9685609023356937,
"eval_rougeLsum": 0.968565755092943,
"eval_runtime": 358.0746,
"eval_samples_per_second": 119.726,
"eval_steps_per_second": 9.978,
"eval_translation_length": 701247,
"step": 230000
},
{
"epoch": 3.61,
"learning_rate": 3.1194509023896597e-06,
"loss": 0.0342,
"step": 232000
},
{
"epoch": 3.64,
"learning_rate": 3.0081077877484786e-06,
"loss": 0.0297,
"step": 234000
},
{
"epoch": 3.67,
"learning_rate": 2.8978719026697843e-06,
"loss": 0.0308,
"step": 236000
},
{
"epoch": 3.7,
"learning_rate": 2.7889177879789993e-06,
"loss": 0.0317,
"step": 238000
},
{
"epoch": 3.73,
"learning_rate": 2.6812000664997108e-06,
"loss": 0.0317,
"step": 240000
},
{
"epoch": 3.73,
"eval_bleu": 0.9545413224765983,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0115161086177766,
"eval_loss": 0.04142594709992409,
"eval_reference_length": 688948,
"eval_rouge1": 0.9706993144139382,
"eval_rouge2": 0.9462893219706924,
"eval_rougeL": 0.9706221623216551,
"eval_rougeLsum": 0.9705871778859747,
"eval_runtime": 352.9472,
"eval_samples_per_second": 121.466,
"eval_steps_per_second": 10.123,
"eval_translation_length": 696882,
"step": 240000
},
{
"epoch": 3.76,
"learning_rate": 2.574942125369937e-06,
"loss": 0.0298,
"step": 242000
},
{
"epoch": 3.79,
"learning_rate": 2.4699932979874153e-06,
"loss": 0.0319,
"step": 244000
},
{
"epoch": 3.83,
"learning_rate": 2.3665732796817735e-06,
"loss": 0.0314,
"step": 246000
},
{
"epoch": 3.86,
"learning_rate": 2.2646389981153643e-06,
"loss": 0.0312,
"step": 248000
},
{
"epoch": 3.89,
"learning_rate": 2.1643020903452345e-06,
"loss": 0.0298,
"step": 250000
},
{
"epoch": 3.89,
"eval_bleu": 0.9529006548919251,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0129792088807863,
"eval_loss": 0.040760207921266556,
"eval_reference_length": 688948,
"eval_rouge1": 0.9693489479213169,
"eval_rouge2": 0.9444261412988928,
"eval_rougeL": 0.9692149741999472,
"eval_rougeLsum": 0.9692295166277185,
"eval_runtime": 355.5559,
"eval_samples_per_second": 120.575,
"eval_steps_per_second": 10.049,
"eval_translation_length": 697890,
"step": 250000
},
{
"epoch": 3.92,
"learning_rate": 2.0656700673466744e-06,
"loss": 0.032,
"step": 252000
},
{
"epoch": 3.95,
"learning_rate": 1.9687498973425523e-06,
"loss": 0.0288,
"step": 254000
},
{
"epoch": 3.98,
"learning_rate": 1.8735981224010946e-06,
"loss": 0.0301,
"step": 256000
},
{
"epoch": 4.01,
"learning_rate": 1.7802702529299903e-06,
"loss": 0.0285,
"step": 258000
},
{
"epoch": 4.04,
"learning_rate": 1.6888207352922886e-06,
"loss": 0.0249,
"step": 260000
},
{
"epoch": 4.04,
"eval_bleu": 0.9580735852306802,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0084215354424426,
"eval_loss": 0.04249930381774902,
"eval_reference_length": 688948,
"eval_rouge1": 0.9725141407368192,
"eval_rouge2": 0.9482526619227644,
"eval_rougeL": 0.9723960508025267,
"eval_rougeLsum": 0.972420107341233,
"eval_runtime": 352.4144,
"eval_samples_per_second": 121.649,
"eval_steps_per_second": 10.139,
"eval_translation_length": 694750,
"step": 260000
},
{
"epoch": 4.07,
"learning_rate": 1.599258630916235e-06,
"loss": 0.0231,
"step": 262000
},
{
"epoch": 4.11,
"learning_rate": 1.5117690308052164e-06,
"loss": 0.0249,
"step": 264000
},
{
"epoch": 4.14,
"learning_rate": 1.4262278806001696e-06,
"loss": 0.0242,
"step": 266000
},
{
"epoch": 4.17,
"learning_rate": 1.3427737126740498e-06,
"loss": 0.0235,
"step": 268000
},
{
"epoch": 4.2,
"learning_rate": 1.261495379512549e-06,
"loss": 0.0236,
"step": 270000
},
{
"epoch": 4.2,
"eval_bleu": 0.9578225644568897,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0083475095362786,
"eval_loss": 0.04301063343882561,
"eval_reference_length": 688948,
"eval_rouge1": 0.9727779850827751,
"eval_rouge2": 0.948452540502593,
"eval_rougeL": 0.9726976320174251,
"eval_rougeLsum": 0.9727011117301205,
"eval_runtime": 360.9368,
"eval_samples_per_second": 118.777,
"eval_steps_per_second": 9.899,
"eval_translation_length": 694699,
"step": 270000
},
{
"epoch": 4.23,
"learning_rate": 1.1823980452421706e-06,
"loss": 0.024,
"step": 272000
},
{
"epoch": 4.26,
"learning_rate": 1.105489964860681e-06,
"loss": 0.0228,
"step": 274000
},
{
"epoch": 4.29,
"learning_rate": 1.0308929099656145e-06,
"loss": 0.0236,
"step": 276000
},
{
"epoch": 4.32,
"learning_rate": 9.586469575615094e-07,
"loss": 0.025,
"step": 278000
},
{
"epoch": 4.35,
"learning_rate": 8.887219403201786e-07,
"loss": 0.0259,
"step": 280000
},
{
"epoch": 4.35,
"eval_bleu": 0.9543405944053881,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0119820363801042,
"eval_loss": 0.04318338260054588,
"eval_reference_length": 688948,
"eval_rouge1": 0.9707381997207458,
"eval_rouge2": 0.9463171031333721,
"eval_rougeL": 0.9706594079415762,
"eval_rougeLsum": 0.9706575791884534,
"eval_runtime": 353.2216,
"eval_samples_per_second": 121.371,
"eval_steps_per_second": 10.115,
"eval_translation_length": 697203,
"step": 280000
},
{
"epoch": 4.39,
"learning_rate": 8.211953935368261e-07,
"loss": 0.0236,
"step": 282000
},
{
"epoch": 4.42,
"learning_rate": 7.560747672425183e-07,
"loss": 0.0243,
"step": 284000
},
{
"epoch": 4.45,
"learning_rate": 6.934324737735693e-07,
"loss": 0.0245,
"step": 286000
},
{
"epoch": 4.48,
"learning_rate": 6.333345238313842e-07,
"loss": 0.0241,
"step": 288000
},
{
"epoch": 4.51,
"learning_rate": 5.757840343120758e-07,
"loss": 0.0236,
"step": 290000
},
{
"epoch": 4.51,
"eval_bleu": 0.9554825445943746,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0108963230896961,
"eval_loss": 0.04305826872587204,
"eval_reference_length": 688948,
"eval_rouge1": 0.9714257759381628,
"eval_rouge2": 0.9469647669616812,
"eval_rougeL": 0.971368415701727,
"eval_rougeLsum": 0.9713609918743913,
"eval_runtime": 360.9244,
"eval_samples_per_second": 118.781,
"eval_steps_per_second": 9.9,
"eval_translation_length": 696455,
"step": 290000
},
{
"epoch": 4.54,
"learning_rate": 5.208145794830483e-07,
"loss": 0.023,
"step": 292000
},
{
"epoch": 4.57,
"learning_rate": 4.684326956861246e-07,
"loss": 0.0248,
"step": 294000
},
{
"epoch": 4.6,
"learning_rate": 4.186971195842365e-07,
"loss": 0.0234,
"step": 296000
},
{
"epoch": 4.63,
"learning_rate": 3.716597523194587e-07,
"loss": 0.023,
"step": 298000
},
{
"epoch": 4.67,
"learning_rate": 3.27322503410189e-07,
"loss": 0.0251,
"step": 300000
},
{
"epoch": 4.67,
"eval_bleu": 0.9555293197464279,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.010778752532847,
"eval_loss": 0.04303622618317604,
"eval_reference_length": 688948,
"eval_rouge1": 0.9711184557484207,
"eval_rouge2": 0.9465872069219559,
"eval_rougeL": 0.971016190214522,
"eval_rougeLsum": 0.9710369956602919,
"eval_runtime": 356.11,
"eval_samples_per_second": 120.387,
"eval_steps_per_second": 10.033,
"eval_translation_length": 696374,
"step": 300000
},
{
"epoch": 4.7,
"learning_rate": 2.857112386772626e-07,
"loss": 0.025,
"step": 302000
},
{
"epoch": 4.73,
"learning_rate": 2.468314855157933e-07,
"loss": 0.0246,
"step": 304000
},
{
"epoch": 4.76,
"learning_rate": 2.1072744891572238e-07,
"loss": 0.0244,
"step": 306000
},
{
"epoch": 4.79,
"learning_rate": 1.7745211624187464e-07,
"loss": 0.0235,
"step": 308000
},
{
"epoch": 4.82,
"learning_rate": 1.469728454791536e-07,
"loss": 0.0241,
"step": 310000
},
{
"epoch": 4.82,
"eval_bleu": 0.9564463934340444,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0099470497047673,
"eval_loss": 0.04298330098390579,
"eval_reference_length": 688948,
"eval_rouge1": 0.9718286385746833,
"eval_rouge2": 0.9474516377136643,
"eval_rougeL": 0.9717457141369523,
"eval_rougeLsum": 0.9717098073885057,
"eval_runtime": 351.9107,
"eval_samples_per_second": 121.824,
"eval_steps_per_second": 10.153,
"eval_translation_length": 695801,
"step": 310000
},
{
"epoch": 4.85,
"learning_rate": 1.1931165304333803e-07,
"loss": 0.0229,
"step": 312000
},
{
"epoch": 4.88,
"learning_rate": 9.450064516007773e-08,
"loss": 0.0245,
"step": 314000
},
{
"epoch": 4.91,
"learning_rate": 7.256456591191674e-08,
"loss": 0.0241,
"step": 316000
},
{
"epoch": 4.94,
"learning_rate": 5.351190850388044e-08,
"loss": 0.0247,
"step": 318000
},
{
"epoch": 4.98,
"learning_rate": 3.7327348382793504e-08,
"loss": 0.025,
"step": 320000
},
{
"epoch": 4.98,
"eval_bleu": 0.9562524667691357,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.010135743191068,
"eval_loss": 0.04290972650051117,
"eval_reference_length": 688948,
"eval_rouge1": 0.9716126931407336,
"eval_rouge2": 0.9473094321074684,
"eval_rougeL": 0.9715449461188315,
"eval_rougeLsum": 0.9715191110303505,
"eval_runtime": 358.3002,
"eval_samples_per_second": 119.651,
"eval_steps_per_second": 9.972,
"eval_translation_length": 695931,
"step": 320000
},
{
"epoch": 5.01,
"learning_rate": 2.4046767073176436e-08,
"loss": 0.0231,
"step": 322000
},
{
"epoch": 5.04,
"learning_rate": 1.3664633482581291e-08,
"loss": 0.0213,
"step": 324000
},
{
"epoch": 5.07,
"learning_rate": 6.200393505542135e-09,
"loss": 0.0229,
"step": 326000
},
{
"epoch": 5.1,
"learning_rate": 1.6450126084593953e-09,
"loss": 0.0226,
"step": 328000
},
{
"epoch": 5.13,
"learning_rate": 5.612241452679357e-12,
"loss": 0.0216,
"step": 330000
},
{
"epoch": 5.13,
"eval_bleu": 0.956286137798364,
"eval_brevity_penalty": 1.0,
"eval_length_ratio": 1.0101197768191503,
"eval_loss": 0.04293430969119072,
"eval_reference_length": 688948,
"eval_rouge1": 0.9716103880266641,
"eval_rouge2": 0.9473106979225123,
"eval_rougeL": 0.9715446246919781,
"eval_rougeLsum": 0.971525437075219,
"eval_runtime": 356.663,
"eval_samples_per_second": 120.2,
"eval_steps_per_second": 10.018,
"eval_translation_length": 695920,
"step": 330000
},
{
"epoch": 5.13,
"step": 330000,
"total_flos": 3.0484200532475904e+16,
"train_loss": 0.08529333166642622,
"train_runtime": 78966.8059,
"train_samples_per_second": 25.074,
"train_steps_per_second": 4.179
}
],
"max_steps": 330000,
"num_train_epochs": 6,
"total_flos": 3.0484200532475904e+16,
"trial_name": null,
"trial_params": null
}