mt5_base-qg-ap-nopeft / trainer_state.json
tiagoblima's picture
End of training
31360a1 verified
{
"best_metric": 1.5832300186157227,
"best_model_checkpoint": "/temp/mt5_base-qg-ap-nopeft/checkpoint-8075",
"epoch": 4.997679096394863,
"eval_steps": 500,
"global_step": 8075,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.06,
"learning_rate": 9.876160990712075e-05,
"loss": 10.6293,
"step": 100
},
{
"epoch": 0.12,
"learning_rate": 9.75232198142415e-05,
"loss": 3.1668,
"step": 200
},
{
"epoch": 0.19,
"learning_rate": 9.628482972136223e-05,
"loss": 2.3399,
"step": 300
},
{
"epoch": 0.25,
"learning_rate": 9.504643962848298e-05,
"loss": 2.2073,
"step": 400
},
{
"epoch": 0.31,
"learning_rate": 9.380804953560372e-05,
"loss": 2.1329,
"step": 500
},
{
"epoch": 0.37,
"learning_rate": 9.256965944272447e-05,
"loss": 2.0981,
"step": 600
},
{
"epoch": 0.43,
"learning_rate": 9.13312693498452e-05,
"loss": 2.0474,
"step": 700
},
{
"epoch": 0.5,
"learning_rate": 9.009287925696595e-05,
"loss": 2.0258,
"step": 800
},
{
"epoch": 0.56,
"learning_rate": 8.885448916408669e-05,
"loss": 1.969,
"step": 900
},
{
"epoch": 0.62,
"learning_rate": 8.761609907120744e-05,
"loss": 1.988,
"step": 1000
},
{
"epoch": 0.68,
"learning_rate": 8.637770897832817e-05,
"loss": 1.9439,
"step": 1100
},
{
"epoch": 0.74,
"learning_rate": 8.513931888544892e-05,
"loss": 1.9251,
"step": 1200
},
{
"epoch": 0.8,
"learning_rate": 8.390092879256966e-05,
"loss": 1.9525,
"step": 1300
},
{
"epoch": 0.87,
"learning_rate": 8.266253869969041e-05,
"loss": 1.9247,
"step": 1400
},
{
"epoch": 0.93,
"learning_rate": 8.142414860681114e-05,
"loss": 1.9087,
"step": 1500
},
{
"epoch": 0.99,
"learning_rate": 8.01857585139319e-05,
"loss": 1.8875,
"step": 1600
},
{
"epoch": 1.0,
"eval_loss": 1.659968614578247,
"eval_runtime": 113.267,
"eval_samples_per_second": 28.764,
"eval_steps_per_second": 7.195,
"step": 1615
},
{
"epoch": 1.05,
"learning_rate": 7.894736842105263e-05,
"loss": 1.7923,
"step": 1700
},
{
"epoch": 1.11,
"learning_rate": 7.770897832817338e-05,
"loss": 1.7924,
"step": 1800
},
{
"epoch": 1.18,
"learning_rate": 7.647058823529411e-05,
"loss": 1.7856,
"step": 1900
},
{
"epoch": 1.24,
"learning_rate": 7.523219814241487e-05,
"loss": 1.8054,
"step": 2000
},
{
"epoch": 1.3,
"learning_rate": 7.39938080495356e-05,
"loss": 1.7597,
"step": 2100
},
{
"epoch": 1.36,
"learning_rate": 7.275541795665635e-05,
"loss": 1.7641,
"step": 2200
},
{
"epoch": 1.42,
"learning_rate": 7.15170278637771e-05,
"loss": 1.7676,
"step": 2300
},
{
"epoch": 1.49,
"learning_rate": 7.027863777089784e-05,
"loss": 1.7533,
"step": 2400
},
{
"epoch": 1.55,
"learning_rate": 6.904024767801857e-05,
"loss": 1.7563,
"step": 2500
},
{
"epoch": 1.61,
"learning_rate": 6.780185758513932e-05,
"loss": 1.7492,
"step": 2600
},
{
"epoch": 1.67,
"learning_rate": 6.656346749226007e-05,
"loss": 1.7344,
"step": 2700
},
{
"epoch": 1.73,
"learning_rate": 6.532507739938081e-05,
"loss": 1.7476,
"step": 2800
},
{
"epoch": 1.79,
"learning_rate": 6.408668730650154e-05,
"loss": 1.7549,
"step": 2900
},
{
"epoch": 1.86,
"learning_rate": 6.28482972136223e-05,
"loss": 1.7315,
"step": 3000
},
{
"epoch": 1.92,
"learning_rate": 6.160990712074304e-05,
"loss": 1.7353,
"step": 3100
},
{
"epoch": 1.98,
"learning_rate": 6.0371517027863775e-05,
"loss": 1.7353,
"step": 3200
},
{
"epoch": 2.0,
"eval_loss": 1.6089082956314087,
"eval_runtime": 113.8532,
"eval_samples_per_second": 28.616,
"eval_steps_per_second": 7.158,
"step": 3231
},
{
"epoch": 2.04,
"learning_rate": 5.913312693498453e-05,
"loss": 1.6808,
"step": 3300
},
{
"epoch": 2.1,
"learning_rate": 5.789473684210527e-05,
"loss": 1.6478,
"step": 3400
},
{
"epoch": 2.17,
"learning_rate": 5.6656346749226006e-05,
"loss": 1.6588,
"step": 3500
},
{
"epoch": 2.23,
"learning_rate": 5.5417956656346745e-05,
"loss": 1.6573,
"step": 3600
},
{
"epoch": 2.29,
"learning_rate": 5.41795665634675e-05,
"loss": 1.6365,
"step": 3700
},
{
"epoch": 2.35,
"learning_rate": 5.294117647058824e-05,
"loss": 1.6457,
"step": 3800
},
{
"epoch": 2.41,
"learning_rate": 5.170278637770898e-05,
"loss": 1.633,
"step": 3900
},
{
"epoch": 2.48,
"learning_rate": 5.046439628482973e-05,
"loss": 1.6463,
"step": 4000
},
{
"epoch": 2.54,
"learning_rate": 4.922600619195047e-05,
"loss": 1.6375,
"step": 4100
},
{
"epoch": 2.6,
"learning_rate": 4.798761609907121e-05,
"loss": 1.6562,
"step": 4200
},
{
"epoch": 2.66,
"learning_rate": 4.6749226006191954e-05,
"loss": 1.6335,
"step": 4300
},
{
"epoch": 2.72,
"learning_rate": 4.551083591331269e-05,
"loss": 1.6242,
"step": 4400
},
{
"epoch": 2.79,
"learning_rate": 4.427244582043344e-05,
"loss": 1.6359,
"step": 4500
},
{
"epoch": 2.85,
"learning_rate": 4.303405572755418e-05,
"loss": 1.6742,
"step": 4600
},
{
"epoch": 2.91,
"learning_rate": 4.1795665634674924e-05,
"loss": 1.6338,
"step": 4700
},
{
"epoch": 2.97,
"learning_rate": 4.055727554179567e-05,
"loss": 1.6382,
"step": 4800
},
{
"epoch": 3.0,
"eval_loss": 1.5925997495651245,
"eval_runtime": 113.9003,
"eval_samples_per_second": 28.604,
"eval_steps_per_second": 7.155,
"step": 4847
},
{
"epoch": 3.03,
"learning_rate": 3.931888544891641e-05,
"loss": 1.6031,
"step": 4900
},
{
"epoch": 3.09,
"learning_rate": 3.8080495356037155e-05,
"loss": 1.5524,
"step": 5000
},
{
"epoch": 3.16,
"learning_rate": 3.6842105263157895e-05,
"loss": 1.5775,
"step": 5100
},
{
"epoch": 3.22,
"learning_rate": 3.560371517027864e-05,
"loss": 1.5465,
"step": 5200
},
{
"epoch": 3.28,
"learning_rate": 3.436532507739939e-05,
"loss": 1.585,
"step": 5300
},
{
"epoch": 3.34,
"learning_rate": 3.3126934984520126e-05,
"loss": 1.5808,
"step": 5400
},
{
"epoch": 3.4,
"learning_rate": 3.188854489164087e-05,
"loss": 1.5785,
"step": 5500
},
{
"epoch": 3.47,
"learning_rate": 3.065015479876161e-05,
"loss": 1.5766,
"step": 5600
},
{
"epoch": 3.53,
"learning_rate": 2.9411764705882354e-05,
"loss": 1.5873,
"step": 5700
},
{
"epoch": 3.59,
"learning_rate": 2.8173374613003096e-05,
"loss": 1.5621,
"step": 5800
},
{
"epoch": 3.65,
"learning_rate": 2.693498452012384e-05,
"loss": 1.5564,
"step": 5900
},
{
"epoch": 3.71,
"learning_rate": 2.5696594427244585e-05,
"loss": 1.5464,
"step": 6000
},
{
"epoch": 3.78,
"learning_rate": 2.4458204334365324e-05,
"loss": 1.5858,
"step": 6100
},
{
"epoch": 3.84,
"learning_rate": 2.3219814241486067e-05,
"loss": 1.5684,
"step": 6200
},
{
"epoch": 3.9,
"learning_rate": 2.1981424148606813e-05,
"loss": 1.5533,
"step": 6300
},
{
"epoch": 3.96,
"learning_rate": 2.0743034055727555e-05,
"loss": 1.5789,
"step": 6400
},
{
"epoch": 4.0,
"eval_loss": 1.5867840051651,
"eval_runtime": 113.9571,
"eval_samples_per_second": 28.59,
"eval_steps_per_second": 7.152,
"step": 6463
},
{
"epoch": 4.02,
"learning_rate": 1.9504643962848298e-05,
"loss": 1.5466,
"step": 6500
},
{
"epoch": 4.08,
"learning_rate": 1.826625386996904e-05,
"loss": 1.5419,
"step": 6600
},
{
"epoch": 4.15,
"learning_rate": 1.7027863777089783e-05,
"loss": 1.5244,
"step": 6700
},
{
"epoch": 4.21,
"learning_rate": 1.5789473684210526e-05,
"loss": 1.5351,
"step": 6800
},
{
"epoch": 4.27,
"learning_rate": 1.4551083591331268e-05,
"loss": 1.5398,
"step": 6900
},
{
"epoch": 4.33,
"learning_rate": 1.3312693498452014e-05,
"loss": 1.5356,
"step": 7000
},
{
"epoch": 4.39,
"learning_rate": 1.2074303405572757e-05,
"loss": 1.5176,
"step": 7100
},
{
"epoch": 4.46,
"learning_rate": 1.08359133126935e-05,
"loss": 1.5288,
"step": 7200
},
{
"epoch": 4.52,
"learning_rate": 9.597523219814242e-06,
"loss": 1.5278,
"step": 7300
},
{
"epoch": 4.58,
"learning_rate": 8.359133126934985e-06,
"loss": 1.5186,
"step": 7400
},
{
"epoch": 4.64,
"learning_rate": 7.120743034055728e-06,
"loss": 1.5089,
"step": 7500
},
{
"epoch": 4.7,
"learning_rate": 5.882352941176471e-06,
"loss": 1.5189,
"step": 7600
},
{
"epoch": 4.77,
"learning_rate": 4.643962848297214e-06,
"loss": 1.5225,
"step": 7700
},
{
"epoch": 4.83,
"learning_rate": 3.4055727554179566e-06,
"loss": 1.5472,
"step": 7800
},
{
"epoch": 4.89,
"learning_rate": 2.1671826625387e-06,
"loss": 1.5072,
"step": 7900
},
{
"epoch": 4.95,
"learning_rate": 9.287925696594427e-07,
"loss": 1.5343,
"step": 8000
},
{
"epoch": 5.0,
"eval_loss": 1.5832300186157227,
"eval_runtime": 113.9636,
"eval_samples_per_second": 28.588,
"eval_steps_per_second": 7.151,
"step": 8075
},
{
"epoch": 5.0,
"step": 8075,
"total_flos": 3.098335204343808e+17,
"train_loss": 1.8253408038726902,
"train_runtime": 36991.002,
"train_samples_per_second": 6.989,
"train_steps_per_second": 0.218
}
],
"logging_steps": 100,
"max_steps": 8075,
"num_train_epochs": 5,
"save_steps": 500,
"total_flos": 3.098335204343808e+17,
"trial_name": null,
"trial_params": null
}