| { |
| "best_global_step": 4000, |
| "best_metric": 4.773771009103065, |
| "best_model_checkpoint": "outputs/bert-tiny-stage2-sbert/checkpoints/checkpoint-4000", |
| "epoch": 0.9460737937559129, |
| "eval_steps": 2000, |
| "global_step": 4000, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.011825922421948912, |
| "grad_norm": 37.789451599121094, |
| "learning_rate": 2.3173327027666118e-07, |
| "loss": 18.0314, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.023651844843897825, |
| "grad_norm": 43.297508239746094, |
| "learning_rate": 4.6819579096713174e-07, |
| "loss": 17.7147, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.035477767265846734, |
| "grad_norm": 36.56124496459961, |
| "learning_rate": 7.046583116576024e-07, |
| "loss": 17.0007, |
| "step": 150 |
| }, |
| { |
| "epoch": 0.04730368968779565, |
| "grad_norm": 34.428916931152344, |
| "learning_rate": 9.411208323480729e-07, |
| "loss": 16.0667, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.05912961210974456, |
| "grad_norm": 33.16645812988281, |
| "learning_rate": 1.1775833530385434e-06, |
| "loss": 14.9131, |
| "step": 250 |
| }, |
| { |
| "epoch": 0.07095553453169347, |
| "grad_norm": 29.064250946044922, |
| "learning_rate": 1.4140458737290142e-06, |
| "loss": 13.9449, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.08278145695364239, |
| "grad_norm": 31.42921257019043, |
| "learning_rate": 1.6505083944194847e-06, |
| "loss": 12.7957, |
| "step": 350 |
| }, |
| { |
| "epoch": 0.0946073793755913, |
| "grad_norm": 33.341365814208984, |
| "learning_rate": 1.8869709151099552e-06, |
| "loss": 11.7288, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.10643330179754021, |
| "grad_norm": 36.34325408935547, |
| "learning_rate": 2.123433435800426e-06, |
| "loss": 10.6945, |
| "step": 450 |
| }, |
| { |
| "epoch": 0.11825922421948912, |
| "grad_norm": 39.06941604614258, |
| "learning_rate": 2.3598959564908965e-06, |
| "loss": 9.4743, |
| "step": 500 |
| }, |
| { |
| "epoch": 0.13008514664143803, |
| "grad_norm": 32.969947814941406, |
| "learning_rate": 2.596358477181367e-06, |
| "loss": 8.6215, |
| "step": 550 |
| }, |
| { |
| "epoch": 0.14191106906338694, |
| "grad_norm": 33.9212760925293, |
| "learning_rate": 2.8328209978718375e-06, |
| "loss": 7.7279, |
| "step": 600 |
| }, |
| { |
| "epoch": 0.15373699148533584, |
| "grad_norm": 32.65876007080078, |
| "learning_rate": 3.069283518562308e-06, |
| "loss": 7.1892, |
| "step": 650 |
| }, |
| { |
| "epoch": 0.16556291390728478, |
| "grad_norm": 29.210859298706055, |
| "learning_rate": 3.3057460392527786e-06, |
| "loss": 6.9682, |
| "step": 700 |
| }, |
| { |
| "epoch": 0.1773888363292337, |
| "grad_norm": 29.231189727783203, |
| "learning_rate": 3.5422085599432495e-06, |
| "loss": 6.4781, |
| "step": 750 |
| }, |
| { |
| "epoch": 0.1892147587511826, |
| "grad_norm": 28.949066162109375, |
| "learning_rate": 3.77867108063372e-06, |
| "loss": 6.1271, |
| "step": 800 |
| }, |
| { |
| "epoch": 0.2010406811731315, |
| "grad_norm": 29.826133728027344, |
| "learning_rate": 4.01513360132419e-06, |
| "loss": 6.1199, |
| "step": 850 |
| }, |
| { |
| "epoch": 0.21286660359508042, |
| "grad_norm": 27.585041046142578, |
| "learning_rate": 4.2515961220146615e-06, |
| "loss": 5.9544, |
| "step": 900 |
| }, |
| { |
| "epoch": 0.22469252601702933, |
| "grad_norm": 28.10279655456543, |
| "learning_rate": 4.488058642705131e-06, |
| "loss": 5.8145, |
| "step": 950 |
| }, |
| { |
| "epoch": 0.23651844843897823, |
| "grad_norm": 26.567943572998047, |
| "learning_rate": 4.7245211633956025e-06, |
| "loss": 5.5599, |
| "step": 1000 |
| }, |
| { |
| "epoch": 0.24834437086092714, |
| "grad_norm": 24.42616081237793, |
| "learning_rate": 4.960983684086072e-06, |
| "loss": 5.2344, |
| "step": 1050 |
| }, |
| { |
| "epoch": 0.26017029328287605, |
| "grad_norm": 25.857810974121094, |
| "learning_rate": 5.197446204776543e-06, |
| "loss": 5.3013, |
| "step": 1100 |
| }, |
| { |
| "epoch": 0.27199621570482496, |
| "grad_norm": 26.047733306884766, |
| "learning_rate": 5.433908725467014e-06, |
| "loss": 5.0562, |
| "step": 1150 |
| }, |
| { |
| "epoch": 0.28382213812677387, |
| "grad_norm": 26.875659942626953, |
| "learning_rate": 5.670371246157485e-06, |
| "loss": 4.8728, |
| "step": 1200 |
| }, |
| { |
| "epoch": 0.2956480605487228, |
| "grad_norm": 21.9539737701416, |
| "learning_rate": 5.906833766847954e-06, |
| "loss": 4.7826, |
| "step": 1250 |
| }, |
| { |
| "epoch": 0.3074739829706717, |
| "grad_norm": 23.06488609313965, |
| "learning_rate": 6.143296287538426e-06, |
| "loss": 4.8806, |
| "step": 1300 |
| }, |
| { |
| "epoch": 0.3192999053926206, |
| "grad_norm": 24.24974250793457, |
| "learning_rate": 6.379758808228896e-06, |
| "loss": 4.6464, |
| "step": 1350 |
| }, |
| { |
| "epoch": 0.33112582781456956, |
| "grad_norm": 22.658571243286133, |
| "learning_rate": 6.616221328919367e-06, |
| "loss": 4.7046, |
| "step": 1400 |
| }, |
| { |
| "epoch": 0.34295175023651847, |
| "grad_norm": 21.927656173706055, |
| "learning_rate": 6.852683849609837e-06, |
| "loss": 4.5188, |
| "step": 1450 |
| }, |
| { |
| "epoch": 0.3547776726584674, |
| "grad_norm": 24.39653778076172, |
| "learning_rate": 7.089146370300309e-06, |
| "loss": 4.4968, |
| "step": 1500 |
| }, |
| { |
| "epoch": 0.3666035950804163, |
| "grad_norm": 23.591333389282227, |
| "learning_rate": 7.325608890990778e-06, |
| "loss": 4.4387, |
| "step": 1550 |
| }, |
| { |
| "epoch": 0.3784295175023652, |
| "grad_norm": 24.572961807250977, |
| "learning_rate": 7.562071411681249e-06, |
| "loss": 4.1702, |
| "step": 1600 |
| }, |
| { |
| "epoch": 0.3902554399243141, |
| "grad_norm": 22.61821174621582, |
| "learning_rate": 7.79853393237172e-06, |
| "loss": 4.2147, |
| "step": 1650 |
| }, |
| { |
| "epoch": 0.402081362346263, |
| "grad_norm": 22.490327835083008, |
| "learning_rate": 8.03499645306219e-06, |
| "loss": 3.9972, |
| "step": 1700 |
| }, |
| { |
| "epoch": 0.4139072847682119, |
| "grad_norm": 23.695873260498047, |
| "learning_rate": 8.271458973752661e-06, |
| "loss": 4.1279, |
| "step": 1750 |
| }, |
| { |
| "epoch": 0.42573320719016083, |
| "grad_norm": 24.085838317871094, |
| "learning_rate": 8.507921494443131e-06, |
| "loss": 4.0214, |
| "step": 1800 |
| }, |
| { |
| "epoch": 0.43755912961210974, |
| "grad_norm": 20.78253173828125, |
| "learning_rate": 8.744384015133602e-06, |
| "loss": 3.9161, |
| "step": 1850 |
| }, |
| { |
| "epoch": 0.44938505203405865, |
| "grad_norm": 19.800090789794922, |
| "learning_rate": 8.980846535824072e-06, |
| "loss": 3.7544, |
| "step": 1900 |
| }, |
| { |
| "epoch": 0.46121097445600756, |
| "grad_norm": 22.900514602661133, |
| "learning_rate": 9.217309056514543e-06, |
| "loss": 3.8246, |
| "step": 1950 |
| }, |
| { |
| "epoch": 0.47303689687795647, |
| "grad_norm": 22.419363021850586, |
| "learning_rate": 9.453771577205015e-06, |
| "loss": 3.7991, |
| "step": 2000 |
| }, |
| { |
| "epoch": 0.47303689687795647, |
| "eval_runtime": 46.7005, |
| "eval_samples_per_second": 0.0, |
| "eval_steps_per_second": 0.0, |
| "eval_validation_loss": 5.98806651504585, |
| "step": 2000 |
| }, |
| { |
| "epoch": 0.4848628192999054, |
| "grad_norm": 22.308876037597656, |
| "learning_rate": 9.690234097895484e-06, |
| "loss": 3.8554, |
| "step": 2050 |
| }, |
| { |
| "epoch": 0.4966887417218543, |
| "grad_norm": 23.8614501953125, |
| "learning_rate": 9.926696618585954e-06, |
| "loss": 3.8123, |
| "step": 2100 |
| }, |
| { |
| "epoch": 0.5085146641438032, |
| "grad_norm": 21.00491714477539, |
| "learning_rate": 1.0163159139276425e-05, |
| "loss": 3.5525, |
| "step": 2150 |
| }, |
| { |
| "epoch": 0.5203405865657521, |
| "grad_norm": 25.555097579956055, |
| "learning_rate": 1.0399621659966897e-05, |
| "loss": 3.5591, |
| "step": 2200 |
| }, |
| { |
| "epoch": 0.532166508987701, |
| "grad_norm": 25.4840087890625, |
| "learning_rate": 1.0636084180657367e-05, |
| "loss": 3.6293, |
| "step": 2250 |
| }, |
| { |
| "epoch": 0.5439924314096499, |
| "grad_norm": 21.117971420288086, |
| "learning_rate": 1.0872546701347836e-05, |
| "loss": 3.5831, |
| "step": 2300 |
| }, |
| { |
| "epoch": 0.5558183538315988, |
| "grad_norm": 23.38995361328125, |
| "learning_rate": 1.1109009222038308e-05, |
| "loss": 3.6007, |
| "step": 2350 |
| }, |
| { |
| "epoch": 0.5676442762535477, |
| "grad_norm": 22.385738372802734, |
| "learning_rate": 1.1345471742728777e-05, |
| "loss": 3.4225, |
| "step": 2400 |
| }, |
| { |
| "epoch": 0.5794701986754967, |
| "grad_norm": 21.53306007385254, |
| "learning_rate": 1.158193426341925e-05, |
| "loss": 3.4405, |
| "step": 2450 |
| }, |
| { |
| "epoch": 0.5912961210974456, |
| "grad_norm": 22.93678092956543, |
| "learning_rate": 1.181839678410972e-05, |
| "loss": 3.4002, |
| "step": 2500 |
| }, |
| { |
| "epoch": 0.6031220435193945, |
| "grad_norm": 20.330045700073242, |
| "learning_rate": 1.2054859304800191e-05, |
| "loss": 3.3653, |
| "step": 2550 |
| }, |
| { |
| "epoch": 0.6149479659413434, |
| "grad_norm": 21.98198699951172, |
| "learning_rate": 1.2291321825490661e-05, |
| "loss": 3.321, |
| "step": 2600 |
| }, |
| { |
| "epoch": 0.6267738883632923, |
| "grad_norm": 18.49015998840332, |
| "learning_rate": 1.252778434618113e-05, |
| "loss": 3.3042, |
| "step": 2650 |
| }, |
| { |
| "epoch": 0.6385998107852412, |
| "grad_norm": 22.69803237915039, |
| "learning_rate": 1.2764246866871602e-05, |
| "loss": 3.2117, |
| "step": 2700 |
| }, |
| { |
| "epoch": 0.6504257332071902, |
| "grad_norm": 19.658132553100586, |
| "learning_rate": 1.3000709387562072e-05, |
| "loss": 3.3423, |
| "step": 2750 |
| }, |
| { |
| "epoch": 0.6622516556291391, |
| "grad_norm": 20.783931732177734, |
| "learning_rate": 1.3237171908252545e-05, |
| "loss": 3.2494, |
| "step": 2800 |
| }, |
| { |
| "epoch": 0.674077578051088, |
| "grad_norm": 17.039609909057617, |
| "learning_rate": 1.3473634428943014e-05, |
| "loss": 3.1364, |
| "step": 2850 |
| }, |
| { |
| "epoch": 0.6859035004730369, |
| "grad_norm": 21.787738800048828, |
| "learning_rate": 1.3710096949633484e-05, |
| "loss": 3.1836, |
| "step": 2900 |
| }, |
| { |
| "epoch": 0.6977294228949859, |
| "grad_norm": 20.883773803710938, |
| "learning_rate": 1.3946559470323956e-05, |
| "loss": 3.1268, |
| "step": 2950 |
| }, |
| { |
| "epoch": 0.7095553453169348, |
| "grad_norm": 17.700597763061523, |
| "learning_rate": 1.4183021991014425e-05, |
| "loss": 3.072, |
| "step": 3000 |
| }, |
| { |
| "epoch": 0.7213812677388837, |
| "grad_norm": 20.23262596130371, |
| "learning_rate": 1.4419484511704895e-05, |
| "loss": 3.0135, |
| "step": 3050 |
| }, |
| { |
| "epoch": 0.7332071901608326, |
| "grad_norm": 19.417842864990234, |
| "learning_rate": 1.4655947032395366e-05, |
| "loss": 3.0607, |
| "step": 3100 |
| }, |
| { |
| "epoch": 0.7450331125827815, |
| "grad_norm": 19.843341827392578, |
| "learning_rate": 1.4892409553085838e-05, |
| "loss": 3.0963, |
| "step": 3150 |
| }, |
| { |
| "epoch": 0.7568590350047304, |
| "grad_norm": 20.248523712158203, |
| "learning_rate": 1.5128872073776309e-05, |
| "loss": 3.0419, |
| "step": 3200 |
| }, |
| { |
| "epoch": 0.7686849574266793, |
| "grad_norm": 24.61260986328125, |
| "learning_rate": 1.5365334594466777e-05, |
| "loss": 2.9891, |
| "step": 3250 |
| }, |
| { |
| "epoch": 0.7805108798486282, |
| "grad_norm": 16.637826919555664, |
| "learning_rate": 1.560179711515725e-05, |
| "loss": 2.9384, |
| "step": 3300 |
| }, |
| { |
| "epoch": 0.7923368022705771, |
| "grad_norm": 24.341026306152344, |
| "learning_rate": 1.583825963584772e-05, |
| "loss": 2.8918, |
| "step": 3350 |
| }, |
| { |
| "epoch": 0.804162724692526, |
| "grad_norm": 18.246440887451172, |
| "learning_rate": 1.607472215653819e-05, |
| "loss": 2.9816, |
| "step": 3400 |
| }, |
| { |
| "epoch": 0.8159886471144749, |
| "grad_norm": 19.296022415161133, |
| "learning_rate": 1.631118467722866e-05, |
| "loss": 2.9664, |
| "step": 3450 |
| }, |
| { |
| "epoch": 0.8278145695364238, |
| "grad_norm": 19.331918716430664, |
| "learning_rate": 1.6547647197919134e-05, |
| "loss": 2.8969, |
| "step": 3500 |
| }, |
| { |
| "epoch": 0.8396404919583728, |
| "grad_norm": 25.586254119873047, |
| "learning_rate": 1.6784109718609602e-05, |
| "loss": 2.9368, |
| "step": 3550 |
| }, |
| { |
| "epoch": 0.8514664143803217, |
| "grad_norm": 19.701223373413086, |
| "learning_rate": 1.7020572239300073e-05, |
| "loss": 2.8513, |
| "step": 3600 |
| }, |
| { |
| "epoch": 0.8632923368022706, |
| "grad_norm": 16.68182945251465, |
| "learning_rate": 1.7257034759990545e-05, |
| "loss": 2.9808, |
| "step": 3650 |
| }, |
| { |
| "epoch": 0.8751182592242195, |
| "grad_norm": 19.592416763305664, |
| "learning_rate": 1.7493497280681013e-05, |
| "loss": 2.8428, |
| "step": 3700 |
| }, |
| { |
| "epoch": 0.8869441816461684, |
| "grad_norm": 20.324504852294922, |
| "learning_rate": 1.7729959801371484e-05, |
| "loss": 2.8775, |
| "step": 3750 |
| }, |
| { |
| "epoch": 0.8987701040681173, |
| "grad_norm": 19.49851417541504, |
| "learning_rate": 1.7966422322061955e-05, |
| "loss": 2.739, |
| "step": 3800 |
| }, |
| { |
| "epoch": 0.9105960264900662, |
| "grad_norm": 19.18546485900879, |
| "learning_rate": 1.8202884842752427e-05, |
| "loss": 2.8277, |
| "step": 3850 |
| }, |
| { |
| "epoch": 0.9224219489120151, |
| "grad_norm": 23.6113338470459, |
| "learning_rate": 1.8439347363442898e-05, |
| "loss": 2.767, |
| "step": 3900 |
| }, |
| { |
| "epoch": 0.934247871333964, |
| "grad_norm": 19.779712677001953, |
| "learning_rate": 1.8675809884133366e-05, |
| "loss": 2.794, |
| "step": 3950 |
| }, |
| { |
| "epoch": 0.9460737937559129, |
| "grad_norm": 23.361425399780273, |
| "learning_rate": 1.8912272404823837e-05, |
| "loss": 2.7738, |
| "step": 4000 |
| }, |
| { |
| "epoch": 0.9460737937559129, |
| "eval_runtime": 47.0317, |
| "eval_samples_per_second": 0.0, |
| "eval_steps_per_second": 0.0, |
| "eval_validation_loss": 4.773771009103065, |
| "step": 4000 |
| } |
| ], |
| "logging_steps": 50, |
| "max_steps": 21140, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 5, |
| "save_steps": 2000, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 0.0, |
| "train_batch_size": 64, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
