| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.14657598499061913, | |
| "eval_steps": 500, | |
| "global_step": 5000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0029315196998123826, | |
| "grad_norm": 0.7248865962028503, | |
| "learning_rate": 4.9868042930033434e-05, | |
| "loss": 1.3593, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.005863039399624765, | |
| "grad_norm": 0.3698488771915436, | |
| "learning_rate": 4.972142396340391e-05, | |
| "loss": 1.3285, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.008794559099437148, | |
| "grad_norm": 0.24848727881908417, | |
| "learning_rate": 4.957480499677439e-05, | |
| "loss": 1.324, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.01172607879924953, | |
| "grad_norm": 0.2648352086544037, | |
| "learning_rate": 4.942818603014486e-05, | |
| "loss": 1.3145, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.014657598499061914, | |
| "grad_norm": 0.3995356559753418, | |
| "learning_rate": 4.928156706351534e-05, | |
| "loss": 1.3141, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.017589118198874296, | |
| "grad_norm": 0.32299748063087463, | |
| "learning_rate": 4.913494809688581e-05, | |
| "loss": 1.3087, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.02052063789868668, | |
| "grad_norm": 0.33237338066101074, | |
| "learning_rate": 4.898832913025629e-05, | |
| "loss": 1.3071, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.02345215759849906, | |
| "grad_norm": 0.2716487646102905, | |
| "learning_rate": 4.8841710163626766e-05, | |
| "loss": 1.3048, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.026383677298311446, | |
| "grad_norm": 0.8654017448425293, | |
| "learning_rate": 4.869509119699725e-05, | |
| "loss": 1.3087, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.02931519699812383, | |
| "grad_norm": 0.27945375442504883, | |
| "learning_rate": 4.854847223036773e-05, | |
| "loss": 1.3015, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.03224671669793621, | |
| "grad_norm": 0.27838173508644104, | |
| "learning_rate": 4.84018532637382e-05, | |
| "loss": 1.2998, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.03517823639774859, | |
| "grad_norm": 0.24157749116420746, | |
| "learning_rate": 4.8255234297108676e-05, | |
| "loss": 1.3016, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.038109756097560975, | |
| "grad_norm": 0.45793822407722473, | |
| "learning_rate": 4.810861533047915e-05, | |
| "loss": 1.2983, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.04104127579737336, | |
| "grad_norm": 0.1312064826488495, | |
| "learning_rate": 4.796199636384963e-05, | |
| "loss": 1.2978, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.04397279549718574, | |
| "grad_norm": 0.33062976598739624, | |
| "learning_rate": 4.7815377397220105e-05, | |
| "loss": 1.2972, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.04690431519699812, | |
| "grad_norm": 0.29819390177726746, | |
| "learning_rate": 4.766875843059058e-05, | |
| "loss": 1.2945, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 0.0498358348968105, | |
| "grad_norm": 0.3396131694316864, | |
| "learning_rate": 4.752213946396106e-05, | |
| "loss": 1.2921, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 0.05276735459662289, | |
| "grad_norm": 0.3231181502342224, | |
| "learning_rate": 4.737552049733154e-05, | |
| "loss": 1.2932, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 0.055698874296435275, | |
| "grad_norm": 0.4015660285949707, | |
| "learning_rate": 4.7228901530702015e-05, | |
| "loss": 1.2899, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 0.05863039399624766, | |
| "grad_norm": 0.436213880777359, | |
| "learning_rate": 4.708228256407249e-05, | |
| "loss": 1.2906, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.06156191369606004, | |
| "grad_norm": 0.3451833426952362, | |
| "learning_rate": 4.693566359744297e-05, | |
| "loss": 1.2884, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 0.06449343339587242, | |
| "grad_norm": 0.41890543699264526, | |
| "learning_rate": 4.6789044630813445e-05, | |
| "loss": 1.2892, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 0.0674249530956848, | |
| "grad_norm": 0.3117181062698364, | |
| "learning_rate": 4.664242566418392e-05, | |
| "loss": 1.2866, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 0.07035647279549719, | |
| "grad_norm": 0.2703840136528015, | |
| "learning_rate": 4.64958066975544e-05, | |
| "loss": 1.2869, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 0.07328799249530957, | |
| "grad_norm": 0.31400740146636963, | |
| "learning_rate": 4.6349187730924874e-05, | |
| "loss": 1.2844, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.07621951219512195, | |
| "grad_norm": 0.36265355348587036, | |
| "learning_rate": 4.620256876429535e-05, | |
| "loss": 1.2847, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 0.07915103189493433, | |
| "grad_norm": 0.32774218916893005, | |
| "learning_rate": 4.605594979766583e-05, | |
| "loss": 1.2889, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 0.08208255159474671, | |
| "grad_norm": 0.26954421401023865, | |
| "learning_rate": 4.590933083103631e-05, | |
| "loss": 1.2823, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 0.0850140712945591, | |
| "grad_norm": 0.4042912423610687, | |
| "learning_rate": 4.5762711864406784e-05, | |
| "loss": 1.2869, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 0.08794559099437148, | |
| "grad_norm": 0.28043991327285767, | |
| "learning_rate": 4.561609289777726e-05, | |
| "loss": 1.2851, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.09087711069418386, | |
| "grad_norm": 0.35115665197372437, | |
| "learning_rate": 4.546947393114774e-05, | |
| "loss": 1.2816, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 0.09380863039399624, | |
| "grad_norm": 0.3325822651386261, | |
| "learning_rate": 4.532285496451821e-05, | |
| "loss": 1.2821, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 0.09674015009380862, | |
| "grad_norm": 0.14176355302333832, | |
| "learning_rate": 4.517623599788869e-05, | |
| "loss": 1.2808, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 0.099671669793621, | |
| "grad_norm": 0.3418448567390442, | |
| "learning_rate": 4.502961703125916e-05, | |
| "loss": 1.2808, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 0.1026031894934334, | |
| "grad_norm": 0.3579727113246918, | |
| "learning_rate": 4.488299806462964e-05, | |
| "loss": 1.2768, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.10553470919324578, | |
| "grad_norm": 0.2470945566892624, | |
| "learning_rate": 4.473637909800012e-05, | |
| "loss": 1.2812, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 0.10846622889305817, | |
| "grad_norm": 0.32134631276130676, | |
| "learning_rate": 4.45897601313706e-05, | |
| "loss": 1.2797, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 0.11139774859287055, | |
| "grad_norm": 0.2982274889945984, | |
| "learning_rate": 4.444314116474108e-05, | |
| "loss": 1.2762, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 0.11432926829268293, | |
| "grad_norm": 0.2519618570804596, | |
| "learning_rate": 4.429652219811155e-05, | |
| "loss": 1.2767, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 0.11726078799249531, | |
| "grad_norm": 0.30841368436813354, | |
| "learning_rate": 4.4149903231482026e-05, | |
| "loss": 1.2765, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.1201923076923077, | |
| "grad_norm": 0.3413415551185608, | |
| "learning_rate": 4.40032842648525e-05, | |
| "loss": 1.274, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 0.12312382739212008, | |
| "grad_norm": 0.3912579119205475, | |
| "learning_rate": 4.385666529822298e-05, | |
| "loss": 1.2694, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 0.12605534709193245, | |
| "grad_norm": 0.39514127373695374, | |
| "learning_rate": 4.3710046331593455e-05, | |
| "loss": 1.2742, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 0.12898686679174484, | |
| "grad_norm": 0.5046316385269165, | |
| "learning_rate": 4.356342736496393e-05, | |
| "loss": 1.2723, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 0.1319183864915572, | |
| "grad_norm": 0.40433749556541443, | |
| "learning_rate": 4.341680839833441e-05, | |
| "loss": 1.2714, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.1348499061913696, | |
| "grad_norm": 0.41701552271842957, | |
| "learning_rate": 4.327018943170489e-05, | |
| "loss": 1.2686, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 0.137781425891182, | |
| "grad_norm": 0.6152161359786987, | |
| "learning_rate": 4.3123570465075365e-05, | |
| "loss": 1.2642, | |
| "step": 4700 | |
| }, | |
| { | |
| "epoch": 0.14071294559099437, | |
| "grad_norm": 0.6579223275184631, | |
| "learning_rate": 4.297695149844584e-05, | |
| "loss": 1.2648, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 0.14364446529080677, | |
| "grad_norm": 0.47039562463760376, | |
| "learning_rate": 4.283033253181632e-05, | |
| "loss": 1.2623, | |
| "step": 4900 | |
| }, | |
| { | |
| "epoch": 0.14657598499061913, | |
| "grad_norm": 0.5053458213806152, | |
| "learning_rate": 4.2683713565186795e-05, | |
| "loss": 1.2581, | |
| "step": 5000 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 34112, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.184673878016e+16, | |
| "train_batch_size": 36, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |