josecannete's picture
adding model finetuned on QA (MLQA)
c51722e
{
"best_metric": 72.54565014579936,
"best_model_checkpoint": "/data/jcanete/all_results/mlqa/albeto_xlarge/epochs_4_bs_16_lr_5e-6/checkpoint-14400",
"epoch": 3.999951250426559,
"global_step": 20512,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.06,
"eval_exact_match": 32.6,
"eval_f1": 56.15794028928574,
"step": 300
},
{
"epoch": 0.1,
"learning_rate": 4.8798264430577225e-06,
"loss": 3.0063,
"step": 500
},
{
"epoch": 0.12,
"eval_exact_match": 38.2,
"eval_f1": 62.74933530055881,
"step": 600
},
{
"epoch": 0.18,
"eval_exact_match": 38.4,
"eval_f1": 64.44399563896586,
"step": 900
},
{
"epoch": 0.19,
"learning_rate": 4.757946567862715e-06,
"loss": 2.1048,
"step": 1000
},
{
"epoch": 0.23,
"eval_exact_match": 39.6,
"eval_f1": 65.51476428813997,
"step": 1200
},
{
"epoch": 0.29,
"learning_rate": 4.6363104524180965e-06,
"loss": 1.9106,
"step": 1500
},
{
"epoch": 0.29,
"eval_exact_match": 39.2,
"eval_f1": 66.56103827722897,
"step": 1500
},
{
"epoch": 0.35,
"eval_exact_match": 40.2,
"eval_f1": 66.52847493412486,
"step": 1800
},
{
"epoch": 0.39,
"learning_rate": 4.51443057722309e-06,
"loss": 1.8845,
"step": 2000
},
{
"epoch": 0.41,
"eval_exact_match": 43.4,
"eval_f1": 68.06136327215253,
"step": 2100
},
{
"epoch": 0.47,
"eval_exact_match": 45.4,
"eval_f1": 68.94645961954767,
"step": 2400
},
{
"epoch": 0.49,
"learning_rate": 4.392550702028082e-06,
"loss": 1.8377,
"step": 2500
},
{
"epoch": 0.53,
"eval_exact_match": 44.6,
"eval_f1": 69.74983503448324,
"step": 2700
},
{
"epoch": 0.58,
"learning_rate": 4.270670826833074e-06,
"loss": 1.8444,
"step": 3000
},
{
"epoch": 0.58,
"eval_exact_match": 44.8,
"eval_f1": 68.72422391765423,
"step": 3000
},
{
"epoch": 0.64,
"eval_exact_match": 44.2,
"eval_f1": 68.80247260343292,
"step": 3300
},
{
"epoch": 0.68,
"learning_rate": 4.1487909516380656e-06,
"loss": 1.7901,
"step": 3500
},
{
"epoch": 0.7,
"eval_exact_match": 45.6,
"eval_f1": 70.11442299691804,
"step": 3600
},
{
"epoch": 0.76,
"eval_exact_match": 45.4,
"eval_f1": 71.84540250423026,
"step": 3900
},
{
"epoch": 0.78,
"learning_rate": 4.026911076443058e-06,
"loss": 1.7667,
"step": 4000
},
{
"epoch": 0.82,
"eval_exact_match": 45.0,
"eval_f1": 71.82625857885415,
"step": 4200
},
{
"epoch": 0.88,
"learning_rate": 3.90503120124805e-06,
"loss": 1.7574,
"step": 4500
},
{
"epoch": 0.88,
"eval_exact_match": 46.2,
"eval_f1": 71.38925591495055,
"step": 4500
},
{
"epoch": 0.94,
"eval_exact_match": 47.0,
"eval_f1": 70.80568844306376,
"step": 4800
},
{
"epoch": 0.97,
"learning_rate": 3.7831513260530424e-06,
"loss": 1.7351,
"step": 5000
},
{
"epoch": 0.99,
"eval_exact_match": 46.4,
"eval_f1": 69.96654581975262,
"step": 5100
},
{
"epoch": 1.05,
"eval_exact_match": 46.8,
"eval_f1": 71.40276694078021,
"step": 5400
},
{
"epoch": 1.07,
"learning_rate": 3.661271450858034e-06,
"loss": 1.5758,
"step": 5500
},
{
"epoch": 1.11,
"eval_exact_match": 45.4,
"eval_f1": 70.95444550622851,
"step": 5700
},
{
"epoch": 1.17,
"learning_rate": 3.539391575663027e-06,
"loss": 1.5533,
"step": 6000
},
{
"epoch": 1.17,
"eval_exact_match": 45.8,
"eval_f1": 71.16796131798688,
"step": 6000
},
{
"epoch": 1.23,
"eval_exact_match": 45.4,
"eval_f1": 71.17415998507276,
"step": 6300
},
{
"epoch": 1.27,
"learning_rate": 3.417511700468019e-06,
"loss": 1.5236,
"step": 6500
},
{
"epoch": 1.29,
"eval_exact_match": 47.2,
"eval_f1": 71.96064560948882,
"step": 6600
},
{
"epoch": 1.35,
"eval_exact_match": 46.8,
"eval_f1": 71.57893095238637,
"step": 6900
},
{
"epoch": 1.37,
"learning_rate": 3.2956318252730114e-06,
"loss": 1.5082,
"step": 7000
},
{
"epoch": 1.4,
"eval_exact_match": 46.2,
"eval_f1": 71.0417637265508,
"step": 7200
},
{
"epoch": 1.46,
"learning_rate": 3.1739957098283936e-06,
"loss": 1.5279,
"step": 7500
},
{
"epoch": 1.46,
"eval_exact_match": 46.8,
"eval_f1": 71.61420929224126,
"step": 7500
},
{
"epoch": 1.52,
"eval_exact_match": 47.4,
"eval_f1": 71.4861929271134,
"step": 7800
},
{
"epoch": 1.56,
"learning_rate": 3.052115834633386e-06,
"loss": 1.5222,
"step": 8000
},
{
"epoch": 1.58,
"eval_exact_match": 45.2,
"eval_f1": 71.80594062526184,
"step": 8100
},
{
"epoch": 1.64,
"eval_exact_match": 47.2,
"eval_f1": 71.47797864113711,
"step": 8400
},
{
"epoch": 1.66,
"learning_rate": 2.9302359594383777e-06,
"loss": 1.5115,
"step": 8500
},
{
"epoch": 1.7,
"eval_exact_match": 46.4,
"eval_f1": 71.83295074421588,
"step": 8700
},
{
"epoch": 1.76,
"learning_rate": 2.80859984399376e-06,
"loss": 1.5468,
"step": 9000
},
{
"epoch": 1.76,
"eval_exact_match": 45.0,
"eval_f1": 71.84617673539766,
"step": 9000
},
{
"epoch": 1.81,
"eval_exact_match": 47.0,
"eval_f1": 72.05451513566037,
"step": 9300
},
{
"epoch": 1.85,
"learning_rate": 2.686719968798752e-06,
"loss": 1.5399,
"step": 9500
},
{
"epoch": 1.87,
"eval_exact_match": 44.8,
"eval_f1": 71.05231817141241,
"step": 9600
},
{
"epoch": 1.93,
"eval_exact_match": 44.8,
"eval_f1": 70.76686017808167,
"step": 9900
},
{
"epoch": 1.95,
"learning_rate": 2.5650838533541343e-06,
"loss": 1.507,
"step": 10000
},
{
"epoch": 1.99,
"eval_exact_match": 45.0,
"eval_f1": 71.0282044654407,
"step": 10200
},
{
"epoch": 2.05,
"learning_rate": 2.4432039781591266e-06,
"loss": 1.4224,
"step": 10500
},
{
"epoch": 2.05,
"eval_exact_match": 45.0,
"eval_f1": 71.18815762834795,
"step": 10500
},
{
"epoch": 2.11,
"eval_exact_match": 45.6,
"eval_f1": 71.5028784465516,
"step": 10800
},
{
"epoch": 2.15,
"learning_rate": 2.321324102964119e-06,
"loss": 1.3113,
"step": 11000
},
{
"epoch": 2.16,
"eval_exact_match": 44.4,
"eval_f1": 71.15086056380413,
"step": 11100
},
{
"epoch": 2.22,
"eval_exact_match": 45.8,
"eval_f1": 71.54431190031305,
"step": 11400
},
{
"epoch": 2.24,
"learning_rate": 2.1994442277691107e-06,
"loss": 1.3522,
"step": 11500
},
{
"epoch": 2.28,
"eval_exact_match": 45.0,
"eval_f1": 71.81209765647928,
"step": 11700
},
{
"epoch": 2.34,
"learning_rate": 2.0775643525741034e-06,
"loss": 1.3144,
"step": 12000
},
{
"epoch": 2.34,
"eval_exact_match": 43.6,
"eval_f1": 71.12699620880493,
"step": 12000
},
{
"epoch": 2.4,
"eval_exact_match": 44.2,
"eval_f1": 71.41750598254089,
"step": 12300
},
{
"epoch": 2.44,
"learning_rate": 1.955684477379095e-06,
"loss": 1.2981,
"step": 12500
},
{
"epoch": 2.46,
"eval_exact_match": 47.2,
"eval_f1": 71.45377213612431,
"step": 12600
},
{
"epoch": 2.52,
"eval_exact_match": 45.6,
"eval_f1": 70.77094339280019,
"step": 12900
},
{
"epoch": 2.54,
"learning_rate": 1.8340483619344776e-06,
"loss": 1.3352,
"step": 13000
},
{
"epoch": 2.57,
"eval_exact_match": 45.2,
"eval_f1": 71.05521519892982,
"step": 13200
},
{
"epoch": 2.63,
"learning_rate": 1.7121684867394696e-06,
"loss": 1.2933,
"step": 13500
},
{
"epoch": 2.63,
"eval_exact_match": 44.0,
"eval_f1": 70.46603727609677,
"step": 13500
},
{
"epoch": 2.69,
"eval_exact_match": 46.2,
"eval_f1": 72.22404409889462,
"step": 13800
},
{
"epoch": 2.73,
"learning_rate": 1.590288611544462e-06,
"loss": 1.3227,
"step": 14000
},
{
"epoch": 2.75,
"eval_exact_match": 44.8,
"eval_f1": 71.47399920055251,
"step": 14100
},
{
"epoch": 2.81,
"eval_exact_match": 47.0,
"eval_f1": 72.54565014579936,
"step": 14400
},
{
"epoch": 2.83,
"learning_rate": 1.468408736349454e-06,
"loss": 1.3303,
"step": 14500
},
{
"epoch": 2.87,
"eval_exact_match": 46.0,
"eval_f1": 71.42400233167065,
"step": 14700
},
{
"epoch": 2.93,
"learning_rate": 1.3467726209048363e-06,
"loss": 1.3055,
"step": 15000
},
{
"epoch": 2.93,
"eval_exact_match": 45.6,
"eval_f1": 71.51695078489598,
"step": 15000
},
{
"epoch": 2.98,
"eval_exact_match": 46.6,
"eval_f1": 71.50435043692885,
"step": 15300
},
{
"epoch": 3.02,
"learning_rate": 1.2248927457098284e-06,
"loss": 1.2986,
"step": 15500
},
{
"epoch": 3.04,
"eval_exact_match": 45.8,
"eval_f1": 71.06499193256445,
"step": 15600
},
{
"epoch": 3.1,
"eval_exact_match": 45.6,
"eval_f1": 70.9580439024025,
"step": 15900
},
{
"epoch": 3.12,
"learning_rate": 1.1032566302652108e-06,
"loss": 1.1564,
"step": 16000
},
{
"epoch": 3.16,
"eval_exact_match": 45.4,
"eval_f1": 70.85970253602206,
"step": 16200
},
{
"epoch": 3.22,
"learning_rate": 9.813767550702028e-07,
"loss": 1.1861,
"step": 16500
},
{
"epoch": 3.22,
"eval_exact_match": 44.8,
"eval_f1": 70.73497821095955,
"step": 16500
},
{
"epoch": 3.28,
"eval_exact_match": 44.4,
"eval_f1": 70.1719318341078,
"step": 16800
},
{
"epoch": 3.32,
"learning_rate": 8.594968798751951e-07,
"loss": 1.177,
"step": 17000
},
{
"epoch": 3.33,
"eval_exact_match": 44.6,
"eval_f1": 71.00348481939916,
"step": 17100
},
{
"epoch": 3.39,
"eval_exact_match": 44.8,
"eval_f1": 71.04230855179065,
"step": 17400
},
{
"epoch": 3.41,
"learning_rate": 7.376170046801873e-07,
"loss": 1.1581,
"step": 17500
},
{
"epoch": 3.45,
"eval_exact_match": 45.0,
"eval_f1": 71.32642226951928,
"step": 17700
},
{
"epoch": 3.51,
"learning_rate": 6.157371294851794e-07,
"loss": 1.1398,
"step": 18000
},
{
"epoch": 3.51,
"eval_exact_match": 44.8,
"eval_f1": 70.45204643207842,
"step": 18000
},
{
"epoch": 3.57,
"eval_exact_match": 44.4,
"eval_f1": 70.37255165848242,
"step": 18300
},
{
"epoch": 3.61,
"learning_rate": 4.938572542901717e-07,
"loss": 1.1478,
"step": 18500
},
{
"epoch": 3.63,
"eval_exact_match": 44.0,
"eval_f1": 70.48540517137832,
"step": 18600
},
{
"epoch": 3.69,
"eval_exact_match": 44.6,
"eval_f1": 70.22760956321906,
"step": 18900
},
{
"epoch": 3.71,
"learning_rate": 3.7197737909516383e-07,
"loss": 1.1574,
"step": 19000
},
{
"epoch": 3.74,
"eval_exact_match": 45.0,
"eval_f1": 70.64118099179048,
"step": 19200
},
{
"epoch": 3.8,
"learning_rate": 2.5009750390015604e-07,
"loss": 1.1667,
"step": 19500
},
{
"epoch": 3.8,
"eval_exact_match": 43.4,
"eval_f1": 70.15239653063759,
"step": 19500
},
{
"epoch": 3.86,
"eval_exact_match": 44.4,
"eval_f1": 70.2883601771379,
"step": 19800
},
{
"epoch": 3.9,
"learning_rate": 1.2846138845553822e-07,
"loss": 1.1602,
"step": 20000
},
{
"epoch": 3.92,
"eval_exact_match": 43.8,
"eval_f1": 70.33719627550457,
"step": 20100
},
{
"epoch": 3.98,
"eval_exact_match": 43.8,
"eval_f1": 70.33111568942401,
"step": 20400
},
{
"epoch": 4.0,
"learning_rate": 6.581513260530422e-09,
"loss": 1.1577,
"step": 20500
},
{
"epoch": 4.0,
"step": 20512,
"total_flos": 2.7267161088888864e+16,
"train_loss": 1.4912065643230206,
"train_runtime": 23906.2096,
"train_samples_per_second": 13.729,
"train_steps_per_second": 0.858
}
],
"max_steps": 20512,
"num_train_epochs": 4,
"total_flos": 2.7267161088888864e+16,
"trial_name": null,
"trial_params": null
}