debias_bert-tiny_kw_mnli_21 / trainer_state.json
jialicheng's picture
Upload folder using huggingface_hub
ef299fb verified
{
"best_metric": 67.98777381558838,
"best_model_checkpoint": "outputs/bert-tiny/kw/mnli_21/checkpoint-61360",
"epoch": 6.0,
"global_step": 73632,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.04,
"learning_rate": 1.9864189482833552e-05,
"loss": 34.5409,
"step": 500
},
{
"epoch": 0.08,
"learning_rate": 1.9728378965667103e-05,
"loss": 32.9539,
"step": 1000
},
{
"epoch": 0.12,
"learning_rate": 1.9592568448500654e-05,
"loss": 31.0086,
"step": 1500
},
{
"epoch": 0.16,
"learning_rate": 1.9456757931334205e-05,
"loss": 30.5966,
"step": 2000
},
{
"epoch": 0.2,
"learning_rate": 1.9320947414167756e-05,
"loss": 30.1725,
"step": 2500
},
{
"epoch": 0.24,
"learning_rate": 1.9185136897001307e-05,
"loss": 29.706,
"step": 3000
},
{
"epoch": 0.29,
"learning_rate": 1.9049326379834858e-05,
"loss": 29.5048,
"step": 3500
},
{
"epoch": 0.33,
"learning_rate": 1.8913515862668405e-05,
"loss": 29.1964,
"step": 4000
},
{
"epoch": 0.37,
"learning_rate": 1.8777705345501956e-05,
"loss": 28.8603,
"step": 4500
},
{
"epoch": 0.41,
"learning_rate": 1.8641894828335507e-05,
"loss": 28.7968,
"step": 5000
},
{
"epoch": 0.45,
"learning_rate": 1.8506084311169058e-05,
"loss": 28.7003,
"step": 5500
},
{
"epoch": 0.49,
"learning_rate": 1.837027379400261e-05,
"loss": 28.4954,
"step": 6000
},
{
"epoch": 0.53,
"learning_rate": 1.823446327683616e-05,
"loss": 28.24,
"step": 6500
},
{
"epoch": 0.57,
"learning_rate": 1.809865275966971e-05,
"loss": 28.1433,
"step": 7000
},
{
"epoch": 0.61,
"learning_rate": 1.796284224250326e-05,
"loss": 27.855,
"step": 7500
},
{
"epoch": 0.65,
"learning_rate": 1.7827031725336812e-05,
"loss": 27.8772,
"step": 8000
},
{
"epoch": 0.69,
"learning_rate": 1.7691221208170363e-05,
"loss": 27.862,
"step": 8500
},
{
"epoch": 0.73,
"learning_rate": 1.7555410691003914e-05,
"loss": 27.767,
"step": 9000
},
{
"epoch": 0.77,
"learning_rate": 1.7419600173837465e-05,
"loss": 27.6884,
"step": 9500
},
{
"epoch": 0.81,
"learning_rate": 1.7283789656671015e-05,
"loss": 27.593,
"step": 10000
},
{
"epoch": 0.86,
"learning_rate": 1.7147979139504566e-05,
"loss": 27.4518,
"step": 10500
},
{
"epoch": 0.9,
"learning_rate": 1.7012168622338114e-05,
"loss": 27.3718,
"step": 11000
},
{
"epoch": 0.94,
"learning_rate": 1.6876358105171665e-05,
"loss": 27.3435,
"step": 11500
},
{
"epoch": 0.98,
"learning_rate": 1.6740547588005215e-05,
"loss": 27.3723,
"step": 12000
},
{
"epoch": 1.0,
"eval_accuracy": 63.87162506367804,
"eval_combined_score": 63.87162506367804,
"eval_loss": 0.8185800313949585,
"eval_runtime": 17.7384,
"eval_samples_per_second": 553.32,
"eval_steps_per_second": 2.199,
"step": 12272
},
{
"epoch": 1.02,
"learning_rate": 1.6604737070838766e-05,
"loss": 27.0058,
"step": 12500
},
{
"epoch": 1.06,
"learning_rate": 1.6468926553672317e-05,
"loss": 26.9092,
"step": 13000
},
{
"epoch": 1.1,
"learning_rate": 1.6333116036505868e-05,
"loss": 26.8185,
"step": 13500
},
{
"epoch": 1.14,
"learning_rate": 1.619730551933942e-05,
"loss": 26.6763,
"step": 14000
},
{
"epoch": 1.18,
"learning_rate": 1.606149500217297e-05,
"loss": 26.7073,
"step": 14500
},
{
"epoch": 1.22,
"learning_rate": 1.592568448500652e-05,
"loss": 26.7014,
"step": 15000
},
{
"epoch": 1.26,
"learning_rate": 1.578987396784007e-05,
"loss": 26.5431,
"step": 15500
},
{
"epoch": 1.3,
"learning_rate": 1.5654063450673622e-05,
"loss": 26.2987,
"step": 16000
},
{
"epoch": 1.34,
"learning_rate": 1.5518252933507173e-05,
"loss": 26.4723,
"step": 16500
},
{
"epoch": 1.39,
"learning_rate": 1.5382442416340724e-05,
"loss": 26.392,
"step": 17000
},
{
"epoch": 1.43,
"learning_rate": 1.5246631899174273e-05,
"loss": 26.1498,
"step": 17500
},
{
"epoch": 1.47,
"learning_rate": 1.5110821382007822e-05,
"loss": 26.1262,
"step": 18000
},
{
"epoch": 1.51,
"learning_rate": 1.4975010864841375e-05,
"loss": 26.1348,
"step": 18500
},
{
"epoch": 1.55,
"learning_rate": 1.4839200347674924e-05,
"loss": 26.0562,
"step": 19000
},
{
"epoch": 1.59,
"learning_rate": 1.4703389830508477e-05,
"loss": 26.2436,
"step": 19500
},
{
"epoch": 1.63,
"learning_rate": 1.4567579313342026e-05,
"loss": 26.1435,
"step": 20000
},
{
"epoch": 1.67,
"learning_rate": 1.4431768796175577e-05,
"loss": 26.1062,
"step": 20500
},
{
"epoch": 1.71,
"learning_rate": 1.4295958279009128e-05,
"loss": 25.9377,
"step": 21000
},
{
"epoch": 1.75,
"learning_rate": 1.4160147761842677e-05,
"loss": 26.3038,
"step": 21500
},
{
"epoch": 1.79,
"learning_rate": 1.402433724467623e-05,
"loss": 26.0764,
"step": 22000
},
{
"epoch": 1.83,
"learning_rate": 1.3888526727509778e-05,
"loss": 26.251,
"step": 22500
},
{
"epoch": 1.87,
"learning_rate": 1.3752716210343331e-05,
"loss": 25.9488,
"step": 23000
},
{
"epoch": 1.91,
"learning_rate": 1.361690569317688e-05,
"loss": 25.7248,
"step": 23500
},
{
"epoch": 1.96,
"learning_rate": 1.3481095176010431e-05,
"loss": 25.5771,
"step": 24000
},
{
"epoch": 2.0,
"learning_rate": 1.3345284658843982e-05,
"loss": 25.9028,
"step": 24500
},
{
"epoch": 2.0,
"eval_accuracy": 65.73611818644932,
"eval_combined_score": 65.73611818644932,
"eval_loss": 0.7796743512153625,
"eval_runtime": 15.909,
"eval_samples_per_second": 616.945,
"eval_steps_per_second": 2.451,
"step": 24544
},
{
"epoch": 2.04,
"learning_rate": 1.3209474141677531e-05,
"loss": 25.5806,
"step": 25000
},
{
"epoch": 2.08,
"learning_rate": 1.3073663624511084e-05,
"loss": 25.5793,
"step": 25500
},
{
"epoch": 2.12,
"learning_rate": 1.2937853107344633e-05,
"loss": 25.0838,
"step": 26000
},
{
"epoch": 2.16,
"learning_rate": 1.2802042590178185e-05,
"loss": 25.2664,
"step": 26500
},
{
"epoch": 2.2,
"learning_rate": 1.2666232073011735e-05,
"loss": 25.4611,
"step": 27000
},
{
"epoch": 2.24,
"learning_rate": 1.2530421555845285e-05,
"loss": 25.336,
"step": 27500
},
{
"epoch": 2.28,
"learning_rate": 1.2394611038678836e-05,
"loss": 25.2676,
"step": 28000
},
{
"epoch": 2.32,
"learning_rate": 1.2258800521512385e-05,
"loss": 25.3788,
"step": 28500
},
{
"epoch": 2.36,
"learning_rate": 1.2122990004345938e-05,
"loss": 25.0569,
"step": 29000
},
{
"epoch": 2.4,
"learning_rate": 1.1987179487179487e-05,
"loss": 25.0311,
"step": 29500
},
{
"epoch": 2.44,
"learning_rate": 1.185136897001304e-05,
"loss": 24.9824,
"step": 30000
},
{
"epoch": 2.49,
"learning_rate": 1.1715558452846589e-05,
"loss": 25.2441,
"step": 30500
},
{
"epoch": 2.53,
"learning_rate": 1.157974793568014e-05,
"loss": 24.9861,
"step": 31000
},
{
"epoch": 2.57,
"learning_rate": 1.144393741851369e-05,
"loss": 24.7679,
"step": 31500
},
{
"epoch": 2.61,
"learning_rate": 1.130812690134724e-05,
"loss": 25.271,
"step": 32000
},
{
"epoch": 2.65,
"learning_rate": 1.1172316384180792e-05,
"loss": 24.9879,
"step": 32500
},
{
"epoch": 2.69,
"learning_rate": 1.1036505867014341e-05,
"loss": 24.9177,
"step": 33000
},
{
"epoch": 2.73,
"learning_rate": 1.0900695349847894e-05,
"loss": 25.384,
"step": 33500
},
{
"epoch": 2.77,
"learning_rate": 1.0764884832681443e-05,
"loss": 25.1108,
"step": 34000
},
{
"epoch": 2.81,
"learning_rate": 1.0629074315514994e-05,
"loss": 24.9601,
"step": 34500
},
{
"epoch": 2.85,
"learning_rate": 1.0493263798348545e-05,
"loss": 24.6376,
"step": 35000
},
{
"epoch": 2.89,
"learning_rate": 1.0357453281182096e-05,
"loss": 25.1261,
"step": 35500
},
{
"epoch": 2.93,
"learning_rate": 1.0221642764015647e-05,
"loss": 24.8459,
"step": 36000
},
{
"epoch": 2.97,
"learning_rate": 1.0085832246849196e-05,
"loss": 25.2014,
"step": 36500
},
{
"epoch": 3.0,
"eval_accuracy": 66.64289353031076,
"eval_combined_score": 66.64289353031076,
"eval_loss": 0.7598663568496704,
"eval_runtime": 15.7255,
"eval_samples_per_second": 624.146,
"eval_steps_per_second": 2.48,
"step": 36816
},
{
"epoch": 3.01,
"learning_rate": 9.950021729682747e-06,
"loss": 24.7148,
"step": 37000
},
{
"epoch": 3.06,
"learning_rate": 9.814211212516298e-06,
"loss": 24.4617,
"step": 37500
},
{
"epoch": 3.1,
"learning_rate": 9.678400695349848e-06,
"loss": 24.7504,
"step": 38000
},
{
"epoch": 3.14,
"learning_rate": 9.5425901781834e-06,
"loss": 24.45,
"step": 38500
},
{
"epoch": 3.18,
"learning_rate": 9.40677966101695e-06,
"loss": 24.6057,
"step": 39000
},
{
"epoch": 3.22,
"learning_rate": 9.2709691438505e-06,
"loss": 24.3787,
"step": 39500
},
{
"epoch": 3.26,
"learning_rate": 9.13515862668405e-06,
"loss": 24.2096,
"step": 40000
},
{
"epoch": 3.3,
"learning_rate": 8.999348109517601e-06,
"loss": 24.5361,
"step": 40500
},
{
"epoch": 3.34,
"learning_rate": 8.863537592351152e-06,
"loss": 24.3345,
"step": 41000
},
{
"epoch": 3.38,
"learning_rate": 8.727727075184703e-06,
"loss": 24.364,
"step": 41500
},
{
"epoch": 3.42,
"learning_rate": 8.591916558018254e-06,
"loss": 24.2865,
"step": 42000
},
{
"epoch": 3.46,
"learning_rate": 8.456106040851804e-06,
"loss": 24.5307,
"step": 42500
},
{
"epoch": 3.5,
"learning_rate": 8.320295523685355e-06,
"loss": 24.5127,
"step": 43000
},
{
"epoch": 3.54,
"learning_rate": 8.184485006518904e-06,
"loss": 24.2719,
"step": 43500
},
{
"epoch": 3.59,
"learning_rate": 8.048674489352455e-06,
"loss": 24.2633,
"step": 44000
},
{
"epoch": 3.63,
"learning_rate": 7.912863972186006e-06,
"loss": 24.1562,
"step": 44500
},
{
"epoch": 3.67,
"learning_rate": 7.777053455019557e-06,
"loss": 24.1874,
"step": 45000
},
{
"epoch": 3.71,
"learning_rate": 7.641242937853108e-06,
"loss": 24.3553,
"step": 45500
},
{
"epoch": 3.75,
"learning_rate": 7.505432420686659e-06,
"loss": 24.549,
"step": 46000
},
{
"epoch": 3.79,
"learning_rate": 7.369621903520209e-06,
"loss": 24.538,
"step": 46500
},
{
"epoch": 3.83,
"learning_rate": 7.23381138635376e-06,
"loss": 24.141,
"step": 47000
},
{
"epoch": 3.87,
"learning_rate": 7.0980008691873105e-06,
"loss": 24.2314,
"step": 47500
},
{
"epoch": 3.91,
"learning_rate": 6.962190352020861e-06,
"loss": 24.7287,
"step": 48000
},
{
"epoch": 3.95,
"learning_rate": 6.826379834854412e-06,
"loss": 24.1801,
"step": 48500
},
{
"epoch": 3.99,
"learning_rate": 6.690569317687963e-06,
"loss": 24.3416,
"step": 49000
},
{
"epoch": 4.0,
"eval_accuracy": 67.66174223127865,
"eval_combined_score": 67.66174223127865,
"eval_loss": 0.7413901090621948,
"eval_runtime": 15.9956,
"eval_samples_per_second": 613.605,
"eval_steps_per_second": 2.438,
"step": 49088
},
{
"epoch": 4.03,
"learning_rate": 6.554758800521513e-06,
"loss": 23.7173,
"step": 49500
},
{
"epoch": 4.07,
"learning_rate": 6.418948283355063e-06,
"loss": 24.159,
"step": 50000
},
{
"epoch": 4.12,
"learning_rate": 6.283137766188614e-06,
"loss": 23.9621,
"step": 50500
},
{
"epoch": 4.16,
"learning_rate": 6.147327249022165e-06,
"loss": 24.1423,
"step": 51000
},
{
"epoch": 4.2,
"learning_rate": 6.011516731855716e-06,
"loss": 24.0207,
"step": 51500
},
{
"epoch": 4.24,
"learning_rate": 5.8757062146892665e-06,
"loss": 24.076,
"step": 52000
},
{
"epoch": 4.28,
"learning_rate": 5.739895697522817e-06,
"loss": 24.0378,
"step": 52500
},
{
"epoch": 4.32,
"learning_rate": 5.6040851803563665e-06,
"loss": 24.0202,
"step": 53000
},
{
"epoch": 4.36,
"learning_rate": 5.468274663189917e-06,
"loss": 24.0662,
"step": 53500
},
{
"epoch": 4.4,
"learning_rate": 5.332464146023468e-06,
"loss": 24.0337,
"step": 54000
},
{
"epoch": 4.44,
"learning_rate": 5.196653628857019e-06,
"loss": 23.7662,
"step": 54500
},
{
"epoch": 4.48,
"learning_rate": 5.06084311169057e-06,
"loss": 23.6851,
"step": 55000
},
{
"epoch": 4.52,
"learning_rate": 4.92503259452412e-06,
"loss": 23.898,
"step": 55500
},
{
"epoch": 4.56,
"learning_rate": 4.789222077357671e-06,
"loss": 23.7756,
"step": 56000
},
{
"epoch": 4.6,
"learning_rate": 4.653411560191222e-06,
"loss": 23.9044,
"step": 56500
},
{
"epoch": 4.64,
"learning_rate": 4.5176010430247726e-06,
"loss": 24.0725,
"step": 57000
},
{
"epoch": 4.69,
"learning_rate": 4.381790525858323e-06,
"loss": 23.7794,
"step": 57500
},
{
"epoch": 4.73,
"learning_rate": 4.2459800086918734e-06,
"loss": 24.1247,
"step": 58000
},
{
"epoch": 4.77,
"learning_rate": 4.110169491525424e-06,
"loss": 24.2018,
"step": 58500
},
{
"epoch": 4.81,
"learning_rate": 3.974358974358974e-06,
"loss": 24.1887,
"step": 59000
},
{
"epoch": 4.85,
"learning_rate": 3.838548457192525e-06,
"loss": 24.0284,
"step": 59500
},
{
"epoch": 4.89,
"learning_rate": 3.702737940026076e-06,
"loss": 23.9692,
"step": 60000
},
{
"epoch": 4.93,
"learning_rate": 3.566927422859627e-06,
"loss": 23.8436,
"step": 60500
},
{
"epoch": 4.97,
"learning_rate": 3.431116905693177e-06,
"loss": 23.7877,
"step": 61000
},
{
"epoch": 5.0,
"eval_accuracy": 67.98777381558838,
"eval_combined_score": 67.98777381558838,
"eval_loss": 0.7376300692558289,
"eval_runtime": 15.7227,
"eval_samples_per_second": 624.258,
"eval_steps_per_second": 2.48,
"step": 61360
},
{
"epoch": 5.01,
"learning_rate": 3.2953063885267278e-06,
"loss": 23.8127,
"step": 61500
},
{
"epoch": 5.05,
"learning_rate": 3.1594958713602786e-06,
"loss": 23.4685,
"step": 62000
},
{
"epoch": 5.09,
"learning_rate": 3.0236853541938286e-06,
"loss": 23.6626,
"step": 62500
},
{
"epoch": 5.13,
"learning_rate": 2.8878748370273795e-06,
"loss": 23.8414,
"step": 63000
},
{
"epoch": 5.17,
"learning_rate": 2.7520643198609304e-06,
"loss": 23.6328,
"step": 63500
},
{
"epoch": 5.22,
"learning_rate": 2.6162538026944812e-06,
"loss": 23.9316,
"step": 64000
},
{
"epoch": 5.26,
"learning_rate": 2.4804432855280312e-06,
"loss": 24.1051,
"step": 64500
},
{
"epoch": 5.3,
"learning_rate": 2.344632768361582e-06,
"loss": 23.7414,
"step": 65000
},
{
"epoch": 5.34,
"learning_rate": 2.2088222511951325e-06,
"loss": 23.8935,
"step": 65500
},
{
"epoch": 5.38,
"learning_rate": 2.0730117340286834e-06,
"loss": 23.6011,
"step": 66000
},
{
"epoch": 5.42,
"learning_rate": 1.937201216862234e-06,
"loss": 23.6583,
"step": 66500
},
{
"epoch": 5.46,
"learning_rate": 1.8013906996957845e-06,
"loss": 23.4516,
"step": 67000
},
{
"epoch": 5.5,
"learning_rate": 1.6655801825293353e-06,
"loss": 23.9621,
"step": 67500
},
{
"epoch": 5.54,
"learning_rate": 1.5297696653628858e-06,
"loss": 23.5286,
"step": 68000
},
{
"epoch": 5.58,
"learning_rate": 1.3939591481964364e-06,
"loss": 23.7543,
"step": 68500
},
{
"epoch": 5.62,
"learning_rate": 1.258148631029987e-06,
"loss": 23.961,
"step": 69000
},
{
"epoch": 5.66,
"learning_rate": 1.1223381138635377e-06,
"loss": 23.4363,
"step": 69500
},
{
"epoch": 5.7,
"learning_rate": 9.865275966970884e-07,
"loss": 23.5779,
"step": 70000
},
{
"epoch": 5.74,
"learning_rate": 8.507170795306389e-07,
"loss": 23.9706,
"step": 70500
},
{
"epoch": 5.79,
"learning_rate": 7.149065623641896e-07,
"loss": 23.6523,
"step": 71000
},
{
"epoch": 5.83,
"learning_rate": 5.790960451977402e-07,
"loss": 23.5014,
"step": 71500
},
{
"epoch": 5.87,
"learning_rate": 4.432855280312908e-07,
"loss": 23.6879,
"step": 72000
},
{
"epoch": 5.91,
"learning_rate": 3.074750108648414e-07,
"loss": 65.8195,
"step": 72500
},
{
"epoch": 5.95,
"learning_rate": 1.7166449369839202e-07,
"loss": 23.5878,
"step": 73000
},
{
"epoch": 5.99,
"learning_rate": 3.585397653194264e-08,
"loss": 23.5938,
"step": 73500
},
{
"epoch": 6.0,
"eval_accuracy": 67.77381558838512,
"eval_combined_score": 67.77381558838512,
"eval_loss": 0.7389720678329468,
"eval_runtime": 15.833,
"eval_samples_per_second": 619.906,
"eval_steps_per_second": 2.463,
"step": 73632
},
{
"epoch": 6.0,
"step": 73632,
"total_flos": 0.0,
"train_loss": 25.68384859034519,
"train_runtime": 8782.0618,
"train_samples_per_second": 268.298,
"train_steps_per_second": 8.384
}
],
"max_steps": 73632,
"num_train_epochs": 6,
"total_flos": 0.0,
"trial_name": null,
"trial_params": null
}