t5_base-qg-ap-oficial / trainer_state.json
tiagoblima's picture
End of training
788c24a verified
raw
history blame
16 kB
{
"best_metric": 1.9945117235183716,
"best_model_checkpoint": "/temp/t5_base-qg-ap-oficial/checkpoint-9466",
"epoch": 4.99894358757659,
"eval_steps": 500,
"global_step": 11830,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.04,
"learning_rate": 9.915469146238378e-05,
"loss": 3.6356,
"step": 100
},
{
"epoch": 0.08,
"learning_rate": 9.830938292476755e-05,
"loss": 2.4983,
"step": 200
},
{
"epoch": 0.13,
"learning_rate": 9.746407438715132e-05,
"loss": 2.4076,
"step": 300
},
{
"epoch": 0.17,
"learning_rate": 9.661876584953509e-05,
"loss": 2.3614,
"step": 400
},
{
"epoch": 0.21,
"learning_rate": 9.577345731191886e-05,
"loss": 2.3141,
"step": 500
},
{
"epoch": 0.25,
"learning_rate": 9.492814877430263e-05,
"loss": 2.3573,
"step": 600
},
{
"epoch": 0.3,
"learning_rate": 9.408284023668639e-05,
"loss": 2.2916,
"step": 700
},
{
"epoch": 0.34,
"learning_rate": 9.323753169907016e-05,
"loss": 2.2743,
"step": 800
},
{
"epoch": 0.38,
"learning_rate": 9.239222316145394e-05,
"loss": 2.2709,
"step": 900
},
{
"epoch": 0.42,
"learning_rate": 9.15469146238377e-05,
"loss": 2.2874,
"step": 1000
},
{
"epoch": 0.46,
"learning_rate": 9.070160608622148e-05,
"loss": 2.2176,
"step": 1100
},
{
"epoch": 0.51,
"learning_rate": 8.985629754860525e-05,
"loss": 2.2363,
"step": 1200
},
{
"epoch": 0.55,
"learning_rate": 8.901098901098901e-05,
"loss": 2.2104,
"step": 1300
},
{
"epoch": 0.59,
"learning_rate": 8.816568047337278e-05,
"loss": 2.2257,
"step": 1400
},
{
"epoch": 0.63,
"learning_rate": 8.732037193575655e-05,
"loss": 2.2096,
"step": 1500
},
{
"epoch": 0.68,
"learning_rate": 8.647506339814032e-05,
"loss": 2.2201,
"step": 1600
},
{
"epoch": 0.72,
"learning_rate": 8.56297548605241e-05,
"loss": 2.1754,
"step": 1700
},
{
"epoch": 0.76,
"learning_rate": 8.478444632290787e-05,
"loss": 2.2049,
"step": 1800
},
{
"epoch": 0.8,
"learning_rate": 8.393913778529164e-05,
"loss": 2.165,
"step": 1900
},
{
"epoch": 0.85,
"learning_rate": 8.309382924767541e-05,
"loss": 2.1974,
"step": 2000
},
{
"epoch": 0.89,
"learning_rate": 8.224852071005918e-05,
"loss": 2.1561,
"step": 2100
},
{
"epoch": 0.93,
"learning_rate": 8.140321217244294e-05,
"loss": 2.1517,
"step": 2200
},
{
"epoch": 0.97,
"learning_rate": 8.055790363482671e-05,
"loss": 2.1593,
"step": 2300
},
{
"epoch": 1.0,
"eval_loss": 2.0508947372436523,
"eval_runtime": 298.019,
"eval_samples_per_second": 35.726,
"eval_steps_per_second": 8.932,
"step": 2366
},
{
"epoch": 1.01,
"learning_rate": 7.971259509721048e-05,
"loss": 2.1273,
"step": 2400
},
{
"epoch": 1.06,
"learning_rate": 7.886728655959425e-05,
"loss": 2.0831,
"step": 2500
},
{
"epoch": 1.1,
"learning_rate": 7.802197802197802e-05,
"loss": 2.0649,
"step": 2600
},
{
"epoch": 1.14,
"learning_rate": 7.71766694843618e-05,
"loss": 2.0674,
"step": 2700
},
{
"epoch": 1.18,
"learning_rate": 7.633136094674557e-05,
"loss": 2.0431,
"step": 2800
},
{
"epoch": 1.23,
"learning_rate": 7.548605240912934e-05,
"loss": 2.051,
"step": 2900
},
{
"epoch": 1.27,
"learning_rate": 7.464074387151311e-05,
"loss": 2.0503,
"step": 3000
},
{
"epoch": 1.31,
"learning_rate": 7.379543533389688e-05,
"loss": 2.0434,
"step": 3100
},
{
"epoch": 1.35,
"learning_rate": 7.295012679628065e-05,
"loss": 2.0142,
"step": 3200
},
{
"epoch": 1.39,
"learning_rate": 7.210481825866442e-05,
"loss": 2.0515,
"step": 3300
},
{
"epoch": 1.44,
"learning_rate": 7.12595097210482e-05,
"loss": 2.0314,
"step": 3400
},
{
"epoch": 1.48,
"learning_rate": 7.041420118343195e-05,
"loss": 2.0394,
"step": 3500
},
{
"epoch": 1.52,
"learning_rate": 6.956889264581573e-05,
"loss": 2.0207,
"step": 3600
},
{
"epoch": 1.56,
"learning_rate": 6.87235841081995e-05,
"loss": 2.0339,
"step": 3700
},
{
"epoch": 1.61,
"learning_rate": 6.787827557058326e-05,
"loss": 2.0312,
"step": 3800
},
{
"epoch": 1.65,
"learning_rate": 6.703296703296703e-05,
"loss": 2.0106,
"step": 3900
},
{
"epoch": 1.69,
"learning_rate": 6.61876584953508e-05,
"loss": 2.03,
"step": 4000
},
{
"epoch": 1.73,
"learning_rate": 6.534234995773457e-05,
"loss": 2.0075,
"step": 4100
},
{
"epoch": 1.77,
"learning_rate": 6.449704142011834e-05,
"loss": 2.0339,
"step": 4200
},
{
"epoch": 1.82,
"learning_rate": 6.365173288250211e-05,
"loss": 2.0053,
"step": 4300
},
{
"epoch": 1.86,
"learning_rate": 6.280642434488588e-05,
"loss": 2.0201,
"step": 4400
},
{
"epoch": 1.9,
"learning_rate": 6.196111580726966e-05,
"loss": 2.0502,
"step": 4500
},
{
"epoch": 1.94,
"learning_rate": 6.111580726965343e-05,
"loss": 1.9973,
"step": 4600
},
{
"epoch": 1.99,
"learning_rate": 6.027049873203719e-05,
"loss": 1.9907,
"step": 4700
},
{
"epoch": 2.0,
"eval_loss": 2.015758514404297,
"eval_runtime": 300.3315,
"eval_samples_per_second": 35.451,
"eval_steps_per_second": 8.864,
"step": 4733
},
{
"epoch": 2.03,
"learning_rate": 5.9425190194420964e-05,
"loss": 1.9456,
"step": 4800
},
{
"epoch": 2.07,
"learning_rate": 5.8579881656804736e-05,
"loss": 1.9231,
"step": 4900
},
{
"epoch": 2.11,
"learning_rate": 5.773457311918851e-05,
"loss": 1.9542,
"step": 5000
},
{
"epoch": 2.16,
"learning_rate": 5.688926458157228e-05,
"loss": 1.9218,
"step": 5100
},
{
"epoch": 2.2,
"learning_rate": 5.604395604395605e-05,
"loss": 1.925,
"step": 5200
},
{
"epoch": 2.24,
"learning_rate": 5.5198647506339815e-05,
"loss": 1.9241,
"step": 5300
},
{
"epoch": 2.28,
"learning_rate": 5.4353338968723586e-05,
"loss": 1.9269,
"step": 5400
},
{
"epoch": 2.32,
"learning_rate": 5.350803043110736e-05,
"loss": 1.9384,
"step": 5500
},
{
"epoch": 2.37,
"learning_rate": 5.266272189349113e-05,
"loss": 1.9325,
"step": 5600
},
{
"epoch": 2.41,
"learning_rate": 5.18174133558749e-05,
"loss": 1.9102,
"step": 5700
},
{
"epoch": 2.45,
"learning_rate": 5.097210481825867e-05,
"loss": 1.9009,
"step": 5800
},
{
"epoch": 2.49,
"learning_rate": 5.012679628064244e-05,
"loss": 1.9148,
"step": 5900
},
{
"epoch": 2.54,
"learning_rate": 4.928148774302621e-05,
"loss": 1.9225,
"step": 6000
},
{
"epoch": 2.58,
"learning_rate": 4.8436179205409974e-05,
"loss": 1.9465,
"step": 6100
},
{
"epoch": 2.62,
"learning_rate": 4.7590870667793745e-05,
"loss": 1.917,
"step": 6200
},
{
"epoch": 2.66,
"learning_rate": 4.674556213017752e-05,
"loss": 1.8996,
"step": 6300
},
{
"epoch": 2.7,
"learning_rate": 4.590025359256129e-05,
"loss": 1.93,
"step": 6400
},
{
"epoch": 2.75,
"learning_rate": 4.505494505494506e-05,
"loss": 1.9288,
"step": 6500
},
{
"epoch": 2.79,
"learning_rate": 4.420963651732883e-05,
"loss": 1.9017,
"step": 6600
},
{
"epoch": 2.83,
"learning_rate": 4.3364327979712596e-05,
"loss": 1.9277,
"step": 6700
},
{
"epoch": 2.87,
"learning_rate": 4.251901944209637e-05,
"loss": 1.9238,
"step": 6800
},
{
"epoch": 2.92,
"learning_rate": 4.167371090448014e-05,
"loss": 1.9105,
"step": 6900
},
{
"epoch": 2.96,
"learning_rate": 4.0828402366863904e-05,
"loss": 1.9408,
"step": 7000
},
{
"epoch": 3.0,
"eval_loss": 2.001091957092285,
"eval_runtime": 301.8989,
"eval_samples_per_second": 35.267,
"eval_steps_per_second": 8.818,
"step": 7099
},
{
"epoch": 3.0,
"learning_rate": 3.9983093829247675e-05,
"loss": 1.8997,
"step": 7100
},
{
"epoch": 3.04,
"learning_rate": 3.913778529163145e-05,
"loss": 1.8823,
"step": 7200
},
{
"epoch": 3.08,
"learning_rate": 3.829247675401522e-05,
"loss": 1.871,
"step": 7300
},
{
"epoch": 3.13,
"learning_rate": 3.744716821639899e-05,
"loss": 1.8499,
"step": 7400
},
{
"epoch": 3.17,
"learning_rate": 3.6601859678782755e-05,
"loss": 1.865,
"step": 7500
},
{
"epoch": 3.21,
"learning_rate": 3.5756551141166526e-05,
"loss": 1.8619,
"step": 7600
},
{
"epoch": 3.25,
"learning_rate": 3.49112426035503e-05,
"loss": 1.8613,
"step": 7700
},
{
"epoch": 3.3,
"learning_rate": 3.406593406593407e-05,
"loss": 1.8476,
"step": 7800
},
{
"epoch": 3.34,
"learning_rate": 3.322062552831784e-05,
"loss": 1.8391,
"step": 7900
},
{
"epoch": 3.38,
"learning_rate": 3.237531699070161e-05,
"loss": 1.8497,
"step": 8000
},
{
"epoch": 3.42,
"learning_rate": 3.153000845308538e-05,
"loss": 1.8666,
"step": 8100
},
{
"epoch": 3.47,
"learning_rate": 3.068469991546914e-05,
"loss": 1.8625,
"step": 8200
},
{
"epoch": 3.51,
"learning_rate": 2.9839391377852917e-05,
"loss": 1.8408,
"step": 8300
},
{
"epoch": 3.55,
"learning_rate": 2.8994082840236685e-05,
"loss": 1.8434,
"step": 8400
},
{
"epoch": 3.59,
"learning_rate": 2.8148774302620457e-05,
"loss": 1.8483,
"step": 8500
},
{
"epoch": 3.63,
"learning_rate": 2.7303465765004228e-05,
"loss": 1.8432,
"step": 8600
},
{
"epoch": 3.68,
"learning_rate": 2.6458157227387996e-05,
"loss": 1.8306,
"step": 8700
},
{
"epoch": 3.72,
"learning_rate": 2.5612848689771768e-05,
"loss": 1.8391,
"step": 8800
},
{
"epoch": 3.76,
"learning_rate": 2.476754015215554e-05,
"loss": 1.8596,
"step": 8900
},
{
"epoch": 3.8,
"learning_rate": 2.3922231614539308e-05,
"loss": 1.8539,
"step": 9000
},
{
"epoch": 3.85,
"learning_rate": 2.307692307692308e-05,
"loss": 1.8408,
"step": 9100
},
{
"epoch": 3.89,
"learning_rate": 2.2231614539306847e-05,
"loss": 1.8436,
"step": 9200
},
{
"epoch": 3.93,
"learning_rate": 2.138630600169062e-05,
"loss": 1.8622,
"step": 9300
},
{
"epoch": 3.97,
"learning_rate": 2.0540997464074387e-05,
"loss": 1.8321,
"step": 9400
},
{
"epoch": 4.0,
"eval_loss": 1.9945117235183716,
"eval_runtime": 300.7616,
"eval_samples_per_second": 35.4,
"eval_steps_per_second": 8.851,
"step": 9466
},
{
"epoch": 4.01,
"learning_rate": 1.969568892645816e-05,
"loss": 1.8354,
"step": 9500
},
{
"epoch": 4.06,
"learning_rate": 1.885038038884193e-05,
"loss": 1.7986,
"step": 9600
},
{
"epoch": 4.1,
"learning_rate": 1.8005071851225698e-05,
"loss": 1.8159,
"step": 9700
},
{
"epoch": 4.14,
"learning_rate": 1.7159763313609466e-05,
"loss": 1.8244,
"step": 9800
},
{
"epoch": 4.18,
"learning_rate": 1.6314454775993238e-05,
"loss": 1.8199,
"step": 9900
},
{
"epoch": 4.23,
"learning_rate": 1.546914623837701e-05,
"loss": 1.8159,
"step": 10000
},
{
"epoch": 4.27,
"learning_rate": 1.462383770076078e-05,
"loss": 1.7997,
"step": 10100
},
{
"epoch": 4.31,
"learning_rate": 1.3778529163144549e-05,
"loss": 1.8197,
"step": 10200
},
{
"epoch": 4.35,
"learning_rate": 1.2933220625528319e-05,
"loss": 1.7971,
"step": 10300
},
{
"epoch": 4.39,
"learning_rate": 1.2087912087912089e-05,
"loss": 1.813,
"step": 10400
},
{
"epoch": 4.44,
"learning_rate": 1.1242603550295859e-05,
"loss": 1.8104,
"step": 10500
},
{
"epoch": 4.48,
"learning_rate": 1.0397295012679628e-05,
"loss": 1.8035,
"step": 10600
},
{
"epoch": 4.52,
"learning_rate": 9.551986475063398e-06,
"loss": 1.8309,
"step": 10700
},
{
"epoch": 4.56,
"learning_rate": 8.70667793744717e-06,
"loss": 1.8331,
"step": 10800
},
{
"epoch": 4.61,
"learning_rate": 7.861369399830938e-06,
"loss": 1.8107,
"step": 10900
},
{
"epoch": 4.65,
"learning_rate": 7.016060862214709e-06,
"loss": 1.803,
"step": 11000
},
{
"epoch": 4.69,
"learning_rate": 6.1707523245984785e-06,
"loss": 1.8064,
"step": 11100
},
{
"epoch": 4.73,
"learning_rate": 5.325443786982249e-06,
"loss": 1.8,
"step": 11200
},
{
"epoch": 4.77,
"learning_rate": 4.480135249366019e-06,
"loss": 1.8006,
"step": 11300
},
{
"epoch": 4.82,
"learning_rate": 3.6348267117497893e-06,
"loss": 1.8271,
"step": 11400
},
{
"epoch": 4.86,
"learning_rate": 2.789518174133559e-06,
"loss": 1.8195,
"step": 11500
},
{
"epoch": 4.9,
"learning_rate": 1.944209636517329e-06,
"loss": 1.8006,
"step": 11600
},
{
"epoch": 4.94,
"learning_rate": 1.098901098901099e-06,
"loss": 1.7734,
"step": 11700
},
{
"epoch": 4.99,
"learning_rate": 2.53592561284869e-07,
"loss": 1.7928,
"step": 11800
},
{
"epoch": 5.0,
"eval_loss": 1.9991737604141235,
"eval_runtime": 302.9112,
"eval_samples_per_second": 35.149,
"eval_steps_per_second": 8.788,
"step": 11830
},
{
"epoch": 5.0,
"step": 11830,
"total_flos": 2.3051248406102016e+17,
"train_loss": 1.9853319837194525,
"train_runtime": 42362.5347,
"train_samples_per_second": 8.937,
"train_steps_per_second": 0.279
}
],
"logging_steps": 100,
"max_steps": 11830,
"num_train_epochs": 5,
"save_steps": 500,
"total_flos": 2.3051248406102016e+17,
"trial_name": null,
"trial_params": null
}