bert-base-cased_mnli_bc / trainer_state.json
mjtwins's picture
commit from mujeen
f1e00c3
{
"best_metric": 0.9330543933054394,
"best_model_checkpoint": "/hdd1/mujeen/retrieval_prf/output/bert-base-cased_mnli_bc/checkpoint-32726",
"epoch": 3.0,
"global_step": 49089,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.03,
"learning_rate": 1.9796695797429162e-05,
"loss": 0.4551,
"step": 500
},
{
"epoch": 0.06,
"learning_rate": 1.959339159485832e-05,
"loss": 0.3461,
"step": 1000
},
{
"epoch": 0.09,
"learning_rate": 1.9389679969035833e-05,
"loss": 0.3143,
"step": 1500
},
{
"epoch": 0.12,
"learning_rate": 1.9185968343213347e-05,
"loss": 0.31,
"step": 2000
},
{
"epoch": 0.15,
"learning_rate": 1.8982256717390864e-05,
"loss": 0.2923,
"step": 2500
},
{
"epoch": 0.18,
"learning_rate": 1.8778545091568378e-05,
"loss": 0.2835,
"step": 3000
},
{
"epoch": 0.21,
"learning_rate": 1.8575240888997535e-05,
"loss": 0.2842,
"step": 3500
},
{
"epoch": 0.24,
"learning_rate": 1.8371529263175052e-05,
"loss": 0.2713,
"step": 4000
},
{
"epoch": 0.28,
"learning_rate": 1.8167817637352566e-05,
"loss": 0.2712,
"step": 4500
},
{
"epoch": 0.31,
"learning_rate": 1.796410601153008e-05,
"loss": 0.256,
"step": 5000
},
{
"epoch": 0.34,
"learning_rate": 1.7760801808959237e-05,
"loss": 0.2602,
"step": 5500
},
{
"epoch": 0.37,
"learning_rate": 1.7557090183136754e-05,
"loss": 0.2619,
"step": 6000
},
{
"epoch": 0.4,
"learning_rate": 1.7353378557314268e-05,
"loss": 0.2528,
"step": 6500
},
{
"epoch": 0.43,
"learning_rate": 1.714966693149178e-05,
"loss": 0.2532,
"step": 7000
},
{
"epoch": 0.46,
"learning_rate": 1.694636272892094e-05,
"loss": 0.2378,
"step": 7500
},
{
"epoch": 0.49,
"learning_rate": 1.6742651103098456e-05,
"loss": 0.242,
"step": 8000
},
{
"epoch": 0.52,
"learning_rate": 1.653893947727597e-05,
"loss": 0.2595,
"step": 8500
},
{
"epoch": 0.55,
"learning_rate": 1.6335227851453483e-05,
"loss": 0.2368,
"step": 9000
},
{
"epoch": 0.58,
"learning_rate": 1.6131923648882644e-05,
"loss": 0.2452,
"step": 9500
},
{
"epoch": 0.61,
"learning_rate": 1.5928212023060157e-05,
"loss": 0.2409,
"step": 10000
},
{
"epoch": 0.64,
"learning_rate": 1.572450039723767e-05,
"loss": 0.2286,
"step": 10500
},
{
"epoch": 0.67,
"learning_rate": 1.5520788771415185e-05,
"loss": 0.2281,
"step": 11000
},
{
"epoch": 0.7,
"learning_rate": 1.5317484568844346e-05,
"loss": 0.2376,
"step": 11500
},
{
"epoch": 0.73,
"learning_rate": 1.5113772943021861e-05,
"loss": 0.228,
"step": 12000
},
{
"epoch": 0.76,
"learning_rate": 1.4910061317199373e-05,
"loss": 0.2257,
"step": 12500
},
{
"epoch": 0.79,
"learning_rate": 1.4706349691376888e-05,
"loss": 0.2328,
"step": 13000
},
{
"epoch": 0.83,
"learning_rate": 1.4502638065554402e-05,
"loss": 0.2375,
"step": 13500
},
{
"epoch": 0.86,
"learning_rate": 1.4298926439731917e-05,
"loss": 0.2255,
"step": 14000
},
{
"epoch": 0.89,
"learning_rate": 1.4095214813909431e-05,
"loss": 0.2217,
"step": 14500
},
{
"epoch": 0.92,
"learning_rate": 1.389191061133859e-05,
"loss": 0.2312,
"step": 15000
},
{
"epoch": 0.95,
"learning_rate": 1.3688198985516104e-05,
"loss": 0.2196,
"step": 15500
},
{
"epoch": 0.98,
"learning_rate": 1.3484894782945264e-05,
"loss": 0.2134,
"step": 16000
},
{
"epoch": 1.0,
"eval_accuracy": 0.9266288105200239,
"eval_loss": 0.21751809120178223,
"eval_runtime": 21.2116,
"eval_samples_per_second": 315.488,
"eval_steps_per_second": 39.46,
"step": 16363
},
{
"epoch": 1.01,
"learning_rate": 1.328118315712278e-05,
"loss": 0.2003,
"step": 16500
},
{
"epoch": 1.04,
"learning_rate": 1.3077471531300292e-05,
"loss": 0.1491,
"step": 17000
},
{
"epoch": 1.07,
"learning_rate": 1.2873759905477807e-05,
"loss": 0.1467,
"step": 17500
},
{
"epoch": 1.1,
"learning_rate": 1.267004827965532e-05,
"loss": 0.1577,
"step": 18000
},
{
"epoch": 1.13,
"learning_rate": 1.2466336653832836e-05,
"loss": 0.152,
"step": 18500
},
{
"epoch": 1.16,
"learning_rate": 1.2263032451261993e-05,
"loss": 0.1622,
"step": 19000
},
{
"epoch": 1.19,
"learning_rate": 1.2059728248691154e-05,
"loss": 0.1485,
"step": 19500
},
{
"epoch": 1.22,
"learning_rate": 1.185601662286867e-05,
"loss": 0.1689,
"step": 20000
},
{
"epoch": 1.25,
"learning_rate": 1.1652304997046182e-05,
"loss": 0.1502,
"step": 20500
},
{
"epoch": 1.28,
"learning_rate": 1.1448593371223697e-05,
"loss": 0.1465,
"step": 21000
},
{
"epoch": 1.31,
"learning_rate": 1.124488174540121e-05,
"loss": 0.1581,
"step": 21500
},
{
"epoch": 1.34,
"learning_rate": 1.1041170119578726e-05,
"loss": 0.1587,
"step": 22000
},
{
"epoch": 1.38,
"learning_rate": 1.083745849375624e-05,
"loss": 0.1479,
"step": 22500
},
{
"epoch": 1.41,
"learning_rate": 1.0633746867933755e-05,
"loss": 0.1601,
"step": 23000
},
{
"epoch": 1.44,
"learning_rate": 1.0430035242111267e-05,
"loss": 0.1529,
"step": 23500
},
{
"epoch": 1.47,
"learning_rate": 1.0226323616288782e-05,
"loss": 0.1596,
"step": 24000
},
{
"epoch": 1.5,
"learning_rate": 1.0022611990466298e-05,
"loss": 0.1505,
"step": 24500
},
{
"epoch": 1.53,
"learning_rate": 9.818900364643811e-06,
"loss": 0.1521,
"step": 25000
},
{
"epoch": 1.56,
"learning_rate": 9.615188738821325e-06,
"loss": 0.16,
"step": 25500
},
{
"epoch": 1.59,
"learning_rate": 9.411477112998839e-06,
"loss": 0.1566,
"step": 26000
},
{
"epoch": 1.62,
"learning_rate": 9.208172910428e-06,
"loss": 0.1518,
"step": 26500
},
{
"epoch": 1.65,
"learning_rate": 9.004461284605513e-06,
"loss": 0.1578,
"step": 27000
},
{
"epoch": 1.68,
"learning_rate": 8.800749658783028e-06,
"loss": 0.154,
"step": 27500
},
{
"epoch": 1.71,
"learning_rate": 8.597038032960542e-06,
"loss": 0.1549,
"step": 28000
},
{
"epoch": 1.74,
"learning_rate": 8.393733830389701e-06,
"loss": 0.1572,
"step": 28500
},
{
"epoch": 1.77,
"learning_rate": 8.190022204567215e-06,
"loss": 0.1458,
"step": 29000
},
{
"epoch": 1.8,
"learning_rate": 7.986310578744728e-06,
"loss": 0.1513,
"step": 29500
},
{
"epoch": 1.83,
"learning_rate": 7.782598952922244e-06,
"loss": 0.1557,
"step": 30000
},
{
"epoch": 1.86,
"learning_rate": 7.578887327099758e-06,
"loss": 0.1467,
"step": 30500
},
{
"epoch": 1.89,
"learning_rate": 7.375583124528917e-06,
"loss": 0.1562,
"step": 31000
},
{
"epoch": 1.93,
"learning_rate": 7.1722789219580764e-06,
"loss": 0.1258,
"step": 31500
},
{
"epoch": 1.96,
"learning_rate": 6.968567296135591e-06,
"loss": 0.1508,
"step": 32000
},
{
"epoch": 1.99,
"learning_rate": 6.7648556703131055e-06,
"loss": 0.1543,
"step": 32500
},
{
"epoch": 2.0,
"eval_accuracy": 0.9330543933054394,
"eval_loss": 0.23859120905399323,
"eval_runtime": 21.7126,
"eval_samples_per_second": 308.209,
"eval_steps_per_second": 38.549,
"step": 32726
},
{
"epoch": 2.02,
"learning_rate": 6.561144044490619e-06,
"loss": 0.1121,
"step": 33000
},
{
"epoch": 2.05,
"learning_rate": 6.357432418668134e-06,
"loss": 0.0824,
"step": 33500
},
{
"epoch": 2.08,
"learning_rate": 6.1541282160972935e-06,
"loss": 0.0791,
"step": 34000
},
{
"epoch": 2.11,
"learning_rate": 5.950416590274808e-06,
"loss": 0.0824,
"step": 34500
},
{
"epoch": 2.14,
"learning_rate": 5.747112387703967e-06,
"loss": 0.0943,
"step": 35000
},
{
"epoch": 2.17,
"learning_rate": 5.543400761881481e-06,
"loss": 0.0827,
"step": 35500
},
{
"epoch": 2.2,
"learning_rate": 5.339689136058995e-06,
"loss": 0.0766,
"step": 36000
},
{
"epoch": 2.23,
"learning_rate": 5.13597751023651e-06,
"loss": 0.0833,
"step": 36500
},
{
"epoch": 2.26,
"learning_rate": 4.9322658844140235e-06,
"loss": 0.0911,
"step": 37000
},
{
"epoch": 2.29,
"learning_rate": 4.728554258591538e-06,
"loss": 0.0921,
"step": 37500
},
{
"epoch": 2.32,
"learning_rate": 4.5248426327690525e-06,
"loss": 0.0978,
"step": 38000
},
{
"epoch": 2.35,
"learning_rate": 4.321131006946567e-06,
"loss": 0.0914,
"step": 38500
},
{
"epoch": 2.38,
"learning_rate": 4.117419381124081e-06,
"loss": 0.0879,
"step": 39000
},
{
"epoch": 2.41,
"learning_rate": 3.913707755301596e-06,
"loss": 0.0755,
"step": 39500
},
{
"epoch": 2.44,
"learning_rate": 3.7104035527307546e-06,
"loss": 0.0926,
"step": 40000
},
{
"epoch": 2.48,
"learning_rate": 3.5066919269082687e-06,
"loss": 0.0859,
"step": 40500
},
{
"epoch": 2.51,
"learning_rate": 3.3029803010857832e-06,
"loss": 0.0956,
"step": 41000
},
{
"epoch": 2.54,
"learning_rate": 3.0992686752632978e-06,
"loss": 0.0927,
"step": 41500
},
{
"epoch": 2.57,
"learning_rate": 2.895557049440812e-06,
"loss": 0.0833,
"step": 42000
},
{
"epoch": 2.6,
"learning_rate": 2.692252846869971e-06,
"loss": 0.0902,
"step": 42500
},
{
"epoch": 2.63,
"learning_rate": 2.4885412210474854e-06,
"loss": 0.0787,
"step": 43000
},
{
"epoch": 2.66,
"learning_rate": 2.2848295952249995e-06,
"loss": 0.0779,
"step": 43500
},
{
"epoch": 2.69,
"learning_rate": 2.081117969402514e-06,
"loss": 0.0739,
"step": 44000
},
{
"epoch": 2.72,
"learning_rate": 1.8774063435800283e-06,
"loss": 0.0818,
"step": 44500
},
{
"epoch": 2.75,
"learning_rate": 1.6736947177575424e-06,
"loss": 0.0886,
"step": 45000
},
{
"epoch": 2.78,
"learning_rate": 1.4703905151867019e-06,
"loss": 0.0852,
"step": 45500
},
{
"epoch": 2.81,
"learning_rate": 1.2666788893642162e-06,
"loss": 0.0766,
"step": 46000
},
{
"epoch": 2.84,
"learning_rate": 1.0629672635417305e-06,
"loss": 0.0786,
"step": 46500
},
{
"epoch": 2.87,
"learning_rate": 8.592556377192448e-07,
"loss": 0.0859,
"step": 47000
},
{
"epoch": 2.9,
"learning_rate": 6.559514351484039e-07,
"loss": 0.0813,
"step": 47500
},
{
"epoch": 2.93,
"learning_rate": 4.5223980932591824e-07,
"loss": 0.0763,
"step": 48000
},
{
"epoch": 2.96,
"learning_rate": 2.4852818350343254e-07,
"loss": 0.0809,
"step": 48500
},
{
"epoch": 2.99,
"learning_rate": 4.481655768094686e-08,
"loss": 0.0913,
"step": 49000
},
{
"epoch": 3.0,
"eval_accuracy": 0.9318589360430365,
"eval_loss": 0.3323625326156616,
"eval_runtime": 19.8853,
"eval_samples_per_second": 336.53,
"eval_steps_per_second": 42.091,
"step": 49089
},
{
"epoch": 3.0,
"step": 49089,
"total_flos": 3.4060342782888e+16,
"train_loss": 0.16549377483758101,
"train_runtime": 5592.5998,
"train_samples_per_second": 140.437,
"train_steps_per_second": 8.777
}
],
"max_steps": 49089,
"num_train_epochs": 3,
"total_flos": 3.4060342782888e+16,
"trial_name": null,
"trial_params": null
}