BERTweet-CoNLL / trainer_state.json
emilys's picture
Upload
db0e4bd
raw
history blame
10.3 kB
{
"best_metric": 0.959878879636639,
"best_model_checkpoint": "trained_models/CoNLL/checkpoint-10000",
"epoch": 0.7121492664862555,
"global_step": 10000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.04,
"learning_rate": 4.821962683378436e-05,
"loss": 0.166,
"step": 500
},
{
"epoch": 0.04,
"eval_accuracy_score": 0.9851641291226977,
"eval_f1": 0.9190357439733999,
"eval_loss": 0.08048456907272339,
"eval_precision": 0.9080157687253614,
"eval_recall": 0.9303264893975093,
"eval_runtime": 27.4476,
"eval_samples_per_second": 118.517,
"eval_steps_per_second": 14.828,
"step": 500
},
{
"epoch": 0.07,
"learning_rate": 4.643925366756873e-05,
"loss": 0.0586,
"step": 1000
},
{
"epoch": 0.07,
"eval_accuracy_score": 0.9874810170943499,
"eval_f1": 0.9253980161707094,
"eval_loss": 0.061091016978025436,
"eval_precision": 0.9167630057803469,
"eval_recall": 0.9341972399865365,
"eval_runtime": 26.4887,
"eval_samples_per_second": 122.807,
"eval_steps_per_second": 15.365,
"step": 1000
},
{
"epoch": 0.11,
"learning_rate": 4.465888050135308e-05,
"loss": 0.0549,
"step": 1500
},
{
"epoch": 0.11,
"eval_accuracy_score": 0.9892138156613839,
"eval_f1": 0.9333000665335994,
"eval_loss": 0.051654551178216934,
"eval_precision": 0.9225583689575797,
"eval_recall": 0.9442948502187816,
"eval_runtime": 27.064,
"eval_samples_per_second": 120.197,
"eval_steps_per_second": 15.038,
"step": 1500
},
{
"epoch": 0.14,
"learning_rate": 4.287850733513745e-05,
"loss": 0.0408,
"step": 2000
},
{
"epoch": 0.14,
"eval_accuracy_score": 0.9887660137845099,
"eval_f1": 0.9369867605161722,
"eval_loss": 0.05154793709516525,
"eval_precision": 0.9330774365821095,
"eval_recall": 0.9409289801413665,
"eval_runtime": 34.9374,
"eval_samples_per_second": 93.109,
"eval_steps_per_second": 11.649,
"step": 2000
},
{
"epoch": 0.18,
"learning_rate": 4.109813416892181e-05,
"loss": 0.0313,
"step": 2500
},
{
"epoch": 0.18,
"eval_accuracy_score": 0.9901094194151319,
"eval_f1": 0.9437919463087249,
"eval_loss": 0.049433596432209015,
"eval_precision": 0.9409501505520241,
"eval_recall": 0.946650959272972,
"eval_runtime": 34.757,
"eval_samples_per_second": 93.593,
"eval_steps_per_second": 11.71,
"step": 2500
},
{
"epoch": 0.21,
"learning_rate": 3.9317761002706174e-05,
"loss": 0.0296,
"step": 3000
},
{
"epoch": 0.21,
"eval_accuracy_score": 0.9906740391729294,
"eval_f1": 0.9458756398422422,
"eval_loss": 0.053381938487291336,
"eval_precision": 0.9432635983263599,
"eval_recall": 0.9485021878155503,
"eval_runtime": 36.167,
"eval_samples_per_second": 89.944,
"eval_steps_per_second": 11.253,
"step": 3000
},
{
"epoch": 0.25,
"learning_rate": 3.7537387836490526e-05,
"loss": 0.0293,
"step": 3500
},
{
"epoch": 0.25,
"eval_accuracy_score": 0.9910634321093416,
"eval_f1": 0.9481096487551345,
"eval_loss": 0.042416807264089584,
"eval_precision": 0.9445465174544847,
"eval_recall": 0.9516997643890945,
"eval_runtime": 34.7402,
"eval_samples_per_second": 93.638,
"eval_steps_per_second": 11.716,
"step": 3500
},
{
"epoch": 0.28,
"learning_rate": 3.575701467027489e-05,
"loss": 0.0226,
"step": 4000
},
{
"epoch": 0.28,
"eval_accuracy_score": 0.9903041158833379,
"eval_f1": 0.9465661641541039,
"eval_loss": 0.04618750512599945,
"eval_precision": 0.9421473824608203,
"eval_recall": 0.9510265903736116,
"eval_runtime": 28.4523,
"eval_samples_per_second": 114.332,
"eval_steps_per_second": 14.305,
"step": 4000
},
{
"epoch": 0.32,
"learning_rate": 3.397664150405925e-05,
"loss": 0.0207,
"step": 4500
},
{
"epoch": 0.32,
"eval_accuracy_score": 0.9913944161052919,
"eval_f1": 0.9520862948407058,
"eval_loss": 0.04386541247367859,
"eval_precision": 0.9461525677247797,
"eval_recall": 0.9580949175361831,
"eval_runtime": 25.9577,
"eval_samples_per_second": 125.319,
"eval_steps_per_second": 15.679,
"step": 4500
},
{
"epoch": 0.36,
"learning_rate": 3.219626833784362e-05,
"loss": 0.0166,
"step": 5000
},
{
"epoch": 0.36,
"eval_accuracy_score": 0.9914528250457537,
"eval_f1": 0.9523649781952366,
"eval_loss": 0.044772420078516006,
"eval_precision": 0.949180875961217,
"eval_recall": 0.9555705149781218,
"eval_runtime": 28.1693,
"eval_samples_per_second": 115.48,
"eval_steps_per_second": 14.448,
"step": 5000
},
{
"epoch": 0.39,
"learning_rate": 3.041589517162797e-05,
"loss": 0.0159,
"step": 5500
},
{
"epoch": 0.39,
"eval_accuracy_score": 0.9914333553989331,
"eval_f1": 0.9543924604510265,
"eval_loss": 0.046620924025774,
"eval_precision": 0.9543924604510265,
"eval_recall": 0.9543924604510265,
"eval_runtime": 27.5426,
"eval_samples_per_second": 118.108,
"eval_steps_per_second": 14.777,
"step": 5500
},
{
"epoch": 0.43,
"learning_rate": 2.8635522005412335e-05,
"loss": 0.0124,
"step": 6000
},
{
"epoch": 0.43,
"eval_accuracy_score": 0.9917254001012422,
"eval_f1": 0.9533227185116904,
"eval_loss": 0.04490247741341591,
"eval_precision": 0.9494241362043064,
"eval_recall": 0.9572534500168294,
"eval_runtime": 22.1337,
"eval_samples_per_second": 146.971,
"eval_steps_per_second": 18.388,
"step": 6000
},
{
"epoch": 0.46,
"learning_rate": 2.6855148839196698e-05,
"loss": 0.0104,
"step": 6500
},
{
"epoch": 0.46,
"eval_accuracy_score": 0.9911607803434446,
"eval_f1": 0.9488426702448842,
"eval_loss": 0.04854311794042587,
"eval_precision": 0.9456703443664326,
"eval_recall": 0.9520363513968361,
"eval_runtime": 22.2143,
"eval_samples_per_second": 146.437,
"eval_steps_per_second": 18.322,
"step": 6500
},
{
"epoch": 0.5,
"learning_rate": 2.507477567298106e-05,
"loss": 0.0116,
"step": 7000
},
{
"epoch": 0.5,
"eval_accuracy_score": 0.9924652466804252,
"eval_f1": 0.958029655692385,
"eval_loss": 0.04380907490849495,
"eval_precision": 0.9537948290241868,
"eval_recall": 0.9623022551329519,
"eval_runtime": 22.1806,
"eval_samples_per_second": 146.659,
"eval_steps_per_second": 18.349,
"step": 7000
},
{
"epoch": 0.53,
"learning_rate": 2.329440250676542e-05,
"loss": 0.008,
"step": 7500
},
{
"epoch": 0.53,
"eval_accuracy_score": 0.9924652466804252,
"eval_f1": 0.9584347972121924,
"eval_loss": 0.046966083347797394,
"eval_precision": 0.9564270152505446,
"eval_recall": 0.9604510265903736,
"eval_runtime": 22.1907,
"eval_samples_per_second": 146.593,
"eval_steps_per_second": 18.341,
"step": 7500
},
{
"epoch": 0.57,
"learning_rate": 2.151402934054978e-05,
"loss": 0.0106,
"step": 8000
},
{
"epoch": 0.57,
"eval_accuracy_score": 0.9920758537440131,
"eval_f1": 0.9552964857837792,
"eval_loss": 0.04950818791985512,
"eval_precision": 0.9521819093797024,
"eval_recall": 0.9584315045439246,
"eval_runtime": 22.1428,
"eval_samples_per_second": 146.91,
"eval_steps_per_second": 18.381,
"step": 8000
},
{
"epoch": 0.61,
"learning_rate": 1.9733656174334143e-05,
"loss": 0.0089,
"step": 8500
},
{
"epoch": 0.61,
"eval_accuracy_score": 0.9921732019781161,
"eval_f1": 0.9558589496607757,
"eval_loss": 0.04179555922746658,
"eval_precision": 0.9514757378689345,
"eval_recall": 0.9602827330865029,
"eval_runtime": 22.1266,
"eval_samples_per_second": 147.018,
"eval_steps_per_second": 18.394,
"step": 8500
},
{
"epoch": 0.64,
"learning_rate": 1.7953283008118502e-05,
"loss": 0.0067,
"step": 9000
},
{
"epoch": 0.64,
"eval_accuracy_score": 0.9919979751567306,
"eval_f1": 0.9558872861455887,
"eval_loss": 0.049222081899642944,
"eval_precision": 0.9526914075560013,
"eval_recall": 0.9591046785594076,
"eval_runtime": 22.1688,
"eval_samples_per_second": 146.738,
"eval_steps_per_second": 18.359,
"step": 9000
},
{
"epoch": 0.68,
"learning_rate": 1.6172909841902865e-05,
"loss": 0.0054,
"step": 9500
},
{
"epoch": 0.68,
"eval_accuracy_score": 0.9919979751567306,
"eval_f1": 0.9572004028197382,
"eval_loss": 0.0546395517885685,
"eval_precision": 0.9546367592902578,
"eval_recall": 0.9597778525748906,
"eval_runtime": 23.9981,
"eval_samples_per_second": 135.552,
"eval_steps_per_second": 16.96,
"step": 9500
},
{
"epoch": 0.71,
"learning_rate": 1.4392536675687223e-05,
"loss": 0.0055,
"step": 10000
},
{
"epoch": 0.71,
"eval_accuracy_score": 0.9924068377399634,
"eval_f1": 0.959878879636639,
"eval_loss": 0.05072605982422829,
"eval_precision": 0.9594753657306205,
"eval_recall": 0.9602827330865029,
"eval_runtime": 22.125,
"eval_samples_per_second": 147.028,
"eval_steps_per_second": 18.396,
"step": 10000
}
],
"max_steps": 14042,
"num_train_epochs": 1,
"total_flos": 5226265866240000.0,
"trial_name": null,
"trial_params": null
}