arabic_xlm_xnli / trainer_state.json
morit's picture
upload new version
697b1cb
{
"best_metric": 0.7437751004016064,
"best_model_checkpoint": "models/arabic_xlm_xnli/checkpoint-49088",
"epoch": 5.0,
"global_step": 61360,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.04,
"learning_rate": 1.983702737940026e-05,
"loss": 0.9951,
"step": 500
},
{
"epoch": 0.08,
"learning_rate": 1.9674054758800523e-05,
"loss": 0.8623,
"step": 1000
},
{
"epoch": 0.12,
"learning_rate": 1.9511082138200782e-05,
"loss": 0.823,
"step": 1500
},
{
"epoch": 0.16,
"learning_rate": 1.9348109517601044e-05,
"loss": 0.7888,
"step": 2000
},
{
"epoch": 0.2,
"learning_rate": 1.9185136897001307e-05,
"loss": 0.7811,
"step": 2500
},
{
"epoch": 0.24,
"learning_rate": 1.9022164276401566e-05,
"loss": 0.7653,
"step": 3000
},
{
"epoch": 0.29,
"learning_rate": 1.8859191655801828e-05,
"loss": 0.7642,
"step": 3500
},
{
"epoch": 0.33,
"learning_rate": 1.8696219035202087e-05,
"loss": 0.742,
"step": 4000
},
{
"epoch": 0.37,
"learning_rate": 1.853324641460235e-05,
"loss": 0.736,
"step": 4500
},
{
"epoch": 0.41,
"learning_rate": 1.837027379400261e-05,
"loss": 0.7308,
"step": 5000
},
{
"epoch": 0.45,
"learning_rate": 1.820730117340287e-05,
"loss": 0.7217,
"step": 5500
},
{
"epoch": 0.49,
"learning_rate": 1.804432855280313e-05,
"loss": 0.719,
"step": 6000
},
{
"epoch": 0.53,
"learning_rate": 1.788135593220339e-05,
"loss": 0.7093,
"step": 6500
},
{
"epoch": 0.57,
"learning_rate": 1.771838331160365e-05,
"loss": 0.7163,
"step": 7000
},
{
"epoch": 0.61,
"learning_rate": 1.7555410691003914e-05,
"loss": 0.7002,
"step": 7500
},
{
"epoch": 0.65,
"learning_rate": 1.7392438070404173e-05,
"loss": 0.6888,
"step": 8000
},
{
"epoch": 0.69,
"learning_rate": 1.7229465449804435e-05,
"loss": 0.6945,
"step": 8500
},
{
"epoch": 0.73,
"learning_rate": 1.7066492829204694e-05,
"loss": 0.6907,
"step": 9000
},
{
"epoch": 0.77,
"learning_rate": 1.6903520208604957e-05,
"loss": 0.6858,
"step": 9500
},
{
"epoch": 0.81,
"learning_rate": 1.6740547588005215e-05,
"loss": 0.6811,
"step": 10000
},
{
"epoch": 0.86,
"learning_rate": 1.6577574967405478e-05,
"loss": 0.678,
"step": 10500
},
{
"epoch": 0.9,
"learning_rate": 1.6414602346805737e-05,
"loss": 0.6778,
"step": 11000
},
{
"epoch": 0.94,
"learning_rate": 1.6251629726206e-05,
"loss": 0.6686,
"step": 11500
},
{
"epoch": 0.98,
"learning_rate": 1.608865710560626e-05,
"loss": 0.6705,
"step": 12000
},
{
"epoch": 1.0,
"eval_accuracy": 0.7152610441767068,
"eval_loss": 0.6905383467674255,
"eval_runtime": 5.4763,
"eval_samples_per_second": 454.685,
"eval_steps_per_second": 56.973,
"step": 12272
},
{
"epoch": 1.02,
"learning_rate": 1.592568448500652e-05,
"loss": 0.6373,
"step": 12500
},
{
"epoch": 1.06,
"learning_rate": 1.576271186440678e-05,
"loss": 0.6081,
"step": 13000
},
{
"epoch": 1.1,
"learning_rate": 1.5599739243807042e-05,
"loss": 0.613,
"step": 13500
},
{
"epoch": 1.14,
"learning_rate": 1.54367666232073e-05,
"loss": 0.6035,
"step": 14000
},
{
"epoch": 1.18,
"learning_rate": 1.5273794002607563e-05,
"loss": 0.6084,
"step": 14500
},
{
"epoch": 1.22,
"learning_rate": 1.5110821382007822e-05,
"loss": 0.6029,
"step": 15000
},
{
"epoch": 1.26,
"learning_rate": 1.4947848761408083e-05,
"loss": 0.5968,
"step": 15500
},
{
"epoch": 1.3,
"learning_rate": 1.4784876140808346e-05,
"loss": 0.5989,
"step": 16000
},
{
"epoch": 1.34,
"learning_rate": 1.4621903520208606e-05,
"loss": 0.6055,
"step": 16500
},
{
"epoch": 1.39,
"learning_rate": 1.4458930899608867e-05,
"loss": 0.6085,
"step": 17000
},
{
"epoch": 1.43,
"learning_rate": 1.4295958279009128e-05,
"loss": 0.5962,
"step": 17500
},
{
"epoch": 1.47,
"learning_rate": 1.4132985658409388e-05,
"loss": 0.5981,
"step": 18000
},
{
"epoch": 1.51,
"learning_rate": 1.3970013037809649e-05,
"loss": 0.6,
"step": 18500
},
{
"epoch": 1.55,
"learning_rate": 1.380704041720991e-05,
"loss": 0.5946,
"step": 19000
},
{
"epoch": 1.59,
"learning_rate": 1.364406779661017e-05,
"loss": 0.6098,
"step": 19500
},
{
"epoch": 1.63,
"learning_rate": 1.3481095176010431e-05,
"loss": 0.5981,
"step": 20000
},
{
"epoch": 1.67,
"learning_rate": 1.3318122555410693e-05,
"loss": 0.5932,
"step": 20500
},
{
"epoch": 1.71,
"learning_rate": 1.3155149934810954e-05,
"loss": 0.6023,
"step": 21000
},
{
"epoch": 1.75,
"learning_rate": 1.2992177314211213e-05,
"loss": 0.6008,
"step": 21500
},
{
"epoch": 1.79,
"learning_rate": 1.2829204693611474e-05,
"loss": 0.5897,
"step": 22000
},
{
"epoch": 1.83,
"learning_rate": 1.2666232073011735e-05,
"loss": 0.5976,
"step": 22500
},
{
"epoch": 1.87,
"learning_rate": 1.2503259452411995e-05,
"loss": 0.5983,
"step": 23000
},
{
"epoch": 1.91,
"learning_rate": 1.2340286831812256e-05,
"loss": 0.5882,
"step": 23500
},
{
"epoch": 1.96,
"learning_rate": 1.2177314211212517e-05,
"loss": 0.5906,
"step": 24000
},
{
"epoch": 2.0,
"learning_rate": 1.2014341590612777e-05,
"loss": 0.5879,
"step": 24500
},
{
"epoch": 2.0,
"eval_accuracy": 0.7421686746987952,
"eval_loss": 0.6315130591392517,
"eval_runtime": 5.2669,
"eval_samples_per_second": 472.767,
"eval_steps_per_second": 59.238,
"step": 24544
},
{
"epoch": 2.04,
"learning_rate": 1.185136897001304e-05,
"loss": 0.5219,
"step": 25000
},
{
"epoch": 2.08,
"learning_rate": 1.16883963494133e-05,
"loss": 0.5187,
"step": 25500
},
{
"epoch": 2.12,
"learning_rate": 1.1525423728813561e-05,
"loss": 0.514,
"step": 26000
},
{
"epoch": 2.16,
"learning_rate": 1.1362451108213822e-05,
"loss": 0.5117,
"step": 26500
},
{
"epoch": 2.2,
"learning_rate": 1.1199478487614082e-05,
"loss": 0.5071,
"step": 27000
},
{
"epoch": 2.24,
"learning_rate": 1.1036505867014341e-05,
"loss": 0.5103,
"step": 27500
},
{
"epoch": 2.28,
"learning_rate": 1.0873533246414602e-05,
"loss": 0.5213,
"step": 28000
},
{
"epoch": 2.32,
"learning_rate": 1.0710560625814863e-05,
"loss": 0.5103,
"step": 28500
},
{
"epoch": 2.36,
"learning_rate": 1.0547588005215125e-05,
"loss": 0.515,
"step": 29000
},
{
"epoch": 2.4,
"learning_rate": 1.0384615384615386e-05,
"loss": 0.5193,
"step": 29500
},
{
"epoch": 2.44,
"learning_rate": 1.0221642764015647e-05,
"loss": 0.5122,
"step": 30000
},
{
"epoch": 2.49,
"learning_rate": 1.0058670143415907e-05,
"loss": 0.5242,
"step": 30500
},
{
"epoch": 2.53,
"learning_rate": 9.895697522816168e-06,
"loss": 0.5206,
"step": 31000
},
{
"epoch": 2.57,
"learning_rate": 9.732724902216429e-06,
"loss": 0.5124,
"step": 31500
},
{
"epoch": 2.61,
"learning_rate": 9.56975228161669e-06,
"loss": 0.5148,
"step": 32000
},
{
"epoch": 2.65,
"learning_rate": 9.40677966101695e-06,
"loss": 0.5075,
"step": 32500
},
{
"epoch": 2.69,
"learning_rate": 9.24380704041721e-06,
"loss": 0.5186,
"step": 33000
},
{
"epoch": 2.73,
"learning_rate": 9.080834419817471e-06,
"loss": 0.5006,
"step": 33500
},
{
"epoch": 2.77,
"learning_rate": 8.917861799217732e-06,
"loss": 0.5253,
"step": 34000
},
{
"epoch": 2.81,
"learning_rate": 8.754889178617993e-06,
"loss": 0.5165,
"step": 34500
},
{
"epoch": 2.85,
"learning_rate": 8.591916558018254e-06,
"loss": 0.5085,
"step": 35000
},
{
"epoch": 2.89,
"learning_rate": 8.428943937418514e-06,
"loss": 0.5157,
"step": 35500
},
{
"epoch": 2.93,
"learning_rate": 8.265971316818775e-06,
"loss": 0.5241,
"step": 36000
},
{
"epoch": 2.97,
"learning_rate": 8.102998696219036e-06,
"loss": 0.5072,
"step": 36500
},
{
"epoch": 3.0,
"eval_accuracy": 0.7409638554216867,
"eval_loss": 0.6780929565429688,
"eval_runtime": 5.9747,
"eval_samples_per_second": 416.756,
"eval_steps_per_second": 52.22,
"step": 36816
},
{
"epoch": 3.01,
"learning_rate": 7.940026075619296e-06,
"loss": 0.4821,
"step": 37000
},
{
"epoch": 3.06,
"learning_rate": 7.777053455019557e-06,
"loss": 0.4348,
"step": 37500
},
{
"epoch": 3.1,
"learning_rate": 7.614080834419818e-06,
"loss": 0.4383,
"step": 38000
},
{
"epoch": 3.14,
"learning_rate": 7.451108213820078e-06,
"loss": 0.4461,
"step": 38500
},
{
"epoch": 3.18,
"learning_rate": 7.288135593220339e-06,
"loss": 0.4436,
"step": 39000
},
{
"epoch": 3.22,
"learning_rate": 7.125162972620601e-06,
"loss": 0.4336,
"step": 39500
},
{
"epoch": 3.26,
"learning_rate": 6.962190352020861e-06,
"loss": 0.4496,
"step": 40000
},
{
"epoch": 3.3,
"learning_rate": 6.799217731421122e-06,
"loss": 0.4466,
"step": 40500
},
{
"epoch": 3.34,
"learning_rate": 6.636245110821382e-06,
"loss": 0.4468,
"step": 41000
},
{
"epoch": 3.38,
"learning_rate": 6.473272490221643e-06,
"loss": 0.4515,
"step": 41500
},
{
"epoch": 3.42,
"learning_rate": 6.310299869621904e-06,
"loss": 0.4377,
"step": 42000
},
{
"epoch": 3.46,
"learning_rate": 6.147327249022165e-06,
"loss": 0.4487,
"step": 42500
},
{
"epoch": 3.5,
"learning_rate": 5.9843546284224255e-06,
"loss": 0.4456,
"step": 43000
},
{
"epoch": 3.54,
"learning_rate": 5.821382007822687e-06,
"loss": 0.4444,
"step": 43500
},
{
"epoch": 3.59,
"learning_rate": 5.658409387222948e-06,
"loss": 0.4431,
"step": 44000
},
{
"epoch": 3.63,
"learning_rate": 5.4954367666232076e-06,
"loss": 0.4561,
"step": 44500
},
{
"epoch": 3.67,
"learning_rate": 5.332464146023468e-06,
"loss": 0.4441,
"step": 45000
},
{
"epoch": 3.71,
"learning_rate": 5.169491525423729e-06,
"loss": 0.4472,
"step": 45500
},
{
"epoch": 3.75,
"learning_rate": 5.0065189048239905e-06,
"loss": 0.4328,
"step": 46000
},
{
"epoch": 3.79,
"learning_rate": 4.843546284224251e-06,
"loss": 0.4353,
"step": 46500
},
{
"epoch": 3.83,
"learning_rate": 4.680573663624511e-06,
"loss": 0.4571,
"step": 47000
},
{
"epoch": 3.87,
"learning_rate": 4.5176010430247726e-06,
"loss": 0.4311,
"step": 47500
},
{
"epoch": 3.91,
"learning_rate": 4.354628422425033e-06,
"loss": 0.4364,
"step": 48000
},
{
"epoch": 3.95,
"learning_rate": 4.191655801825294e-06,
"loss": 0.4363,
"step": 48500
},
{
"epoch": 3.99,
"learning_rate": 4.028683181225555e-06,
"loss": 0.4331,
"step": 49000
},
{
"epoch": 4.0,
"eval_accuracy": 0.7437751004016064,
"eval_loss": 0.716753363609314,
"eval_runtime": 5.7541,
"eval_samples_per_second": 432.737,
"eval_steps_per_second": 54.222,
"step": 49088
},
{
"epoch": 4.03,
"learning_rate": 3.865710560625815e-06,
"loss": 0.3925,
"step": 49500
},
{
"epoch": 4.07,
"learning_rate": 3.702737940026076e-06,
"loss": 0.3896,
"step": 50000
},
{
"epoch": 4.12,
"learning_rate": 3.5397653194263363e-06,
"loss": 0.3936,
"step": 50500
},
{
"epoch": 4.16,
"learning_rate": 3.3767926988265974e-06,
"loss": 0.3882,
"step": 51000
},
{
"epoch": 4.2,
"learning_rate": 3.213820078226858e-06,
"loss": 0.38,
"step": 51500
},
{
"epoch": 4.24,
"learning_rate": 3.0508474576271192e-06,
"loss": 0.3725,
"step": 52000
},
{
"epoch": 4.28,
"learning_rate": 2.8878748370273795e-06,
"loss": 0.3944,
"step": 52500
},
{
"epoch": 4.32,
"learning_rate": 2.7249022164276406e-06,
"loss": 0.3769,
"step": 53000
},
{
"epoch": 4.36,
"learning_rate": 2.5619295958279013e-06,
"loss": 0.3729,
"step": 53500
},
{
"epoch": 4.4,
"learning_rate": 2.398956975228162e-06,
"loss": 0.3885,
"step": 54000
},
{
"epoch": 4.44,
"learning_rate": 2.2359843546284227e-06,
"loss": 0.3964,
"step": 54500
},
{
"epoch": 4.48,
"learning_rate": 2.0730117340286834e-06,
"loss": 0.3946,
"step": 55000
},
{
"epoch": 4.52,
"learning_rate": 1.910039113428944e-06,
"loss": 0.3922,
"step": 55500
},
{
"epoch": 4.56,
"learning_rate": 1.7470664928292048e-06,
"loss": 0.3811,
"step": 56000
},
{
"epoch": 4.6,
"learning_rate": 1.5840938722294655e-06,
"loss": 0.3892,
"step": 56500
},
{
"epoch": 4.64,
"learning_rate": 1.4211212516297262e-06,
"loss": 0.388,
"step": 57000
},
{
"epoch": 4.69,
"learning_rate": 1.258148631029987e-06,
"loss": 0.3793,
"step": 57500
},
{
"epoch": 4.73,
"learning_rate": 1.0951760104302478e-06,
"loss": 0.3826,
"step": 58000
},
{
"epoch": 4.77,
"learning_rate": 9.322033898305086e-07,
"loss": 0.3843,
"step": 58500
},
{
"epoch": 4.81,
"learning_rate": 7.692307692307694e-07,
"loss": 0.3711,
"step": 59000
},
{
"epoch": 4.85,
"learning_rate": 6.0625814863103e-07,
"loss": 0.3954,
"step": 59500
},
{
"epoch": 4.89,
"learning_rate": 4.432855280312908e-07,
"loss": 0.3805,
"step": 60000
},
{
"epoch": 4.93,
"learning_rate": 2.803129074315515e-07,
"loss": 0.3904,
"step": 60500
},
{
"epoch": 4.97,
"learning_rate": 1.1734028683181226e-07,
"loss": 0.3885,
"step": 61000
},
{
"epoch": 5.0,
"eval_accuracy": 0.7373493975903614,
"eval_loss": 0.7633547782897949,
"eval_runtime": 5.7535,
"eval_samples_per_second": 432.782,
"eval_steps_per_second": 54.228,
"step": 61360
},
{
"epoch": 5.0,
"step": 61360,
"total_flos": 1.2915645671077632e+17,
"train_loss": 0.5355199811356008,
"train_runtime": 12222.1805,
"train_samples_per_second": 160.651,
"train_steps_per_second": 5.02
}
],
"max_steps": 61360,
"num_train_epochs": 5,
"total_flos": 1.2915645671077632e+17,
"trial_name": null,
"trial_params": null
}