| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 1.0, |
| "eval_steps": 500, |
| "global_step": 268, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.018656716417910446, |
| "grad_norm": 4.868363004512183, |
| "learning_rate": 1.785714285714286e-05, |
| "loss": 0.785, |
| "step": 5 |
| }, |
| { |
| "epoch": 0.03731343283582089, |
| "grad_norm": 1.751933509835608, |
| "learning_rate": 3.571428571428572e-05, |
| "loss": 0.6716, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.055970149253731345, |
| "grad_norm": 81.07935807927798, |
| "learning_rate": 4.999827900623038e-05, |
| "loss": 0.6529, |
| "step": 15 |
| }, |
| { |
| "epoch": 0.07462686567164178, |
| "grad_norm": 2.8797972768135103, |
| "learning_rate": 4.993807186343243e-05, |
| "loss": 0.9759, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.09328358208955224, |
| "grad_norm": 1.7163774373722742, |
| "learning_rate": 4.979207812402531e-05, |
| "loss": 0.6153, |
| "step": 25 |
| }, |
| { |
| "epoch": 0.11194029850746269, |
| "grad_norm": 1.0647139467888829, |
| "learning_rate": 4.956085596012407e-05, |
| "loss": 0.5524, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.13059701492537312, |
| "grad_norm": 0.6190108052239293, |
| "learning_rate": 4.924528939432311e-05, |
| "loss": 0.5182, |
| "step": 35 |
| }, |
| { |
| "epoch": 0.14925373134328357, |
| "grad_norm": 0.9072546858829459, |
| "learning_rate": 4.884658491984735e-05, |
| "loss": 0.5015, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.16791044776119404, |
| "grad_norm": 1.03962717041212, |
| "learning_rate": 4.8366266887814235e-05, |
| "loss": 0.4989, |
| "step": 45 |
| }, |
| { |
| "epoch": 0.1865671641791045, |
| "grad_norm": 0.6168260303205843, |
| "learning_rate": 4.780617167924209e-05, |
| "loss": 0.4799, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.20522388059701493, |
| "grad_norm": 0.7135148450881054, |
| "learning_rate": 4.716844068408693e-05, |
| "loss": 0.478, |
| "step": 55 |
| }, |
| { |
| "epoch": 0.22388059701492538, |
| "grad_norm": 0.5289887894098283, |
| "learning_rate": 4.6455512114150546e-05, |
| "loss": 0.4724, |
| "step": 60 |
| }, |
| { |
| "epoch": 0.24253731343283583, |
| "grad_norm": 0.7621147273264112, |
| "learning_rate": 4.5670111681161296e-05, |
| "loss": 0.4633, |
| "step": 65 |
| }, |
| { |
| "epoch": 0.26119402985074625, |
| "grad_norm": 0.6668759994988325, |
| "learning_rate": 4.481524217566783e-05, |
| "loss": 0.4587, |
| "step": 70 |
| }, |
| { |
| "epoch": 0.2798507462686567, |
| "grad_norm": 0.8149740115435178, |
| "learning_rate": 4.3894171986588217e-05, |
| "loss": 0.4576, |
| "step": 75 |
| }, |
| { |
| "epoch": 0.29850746268656714, |
| "grad_norm": 0.7620741688125229, |
| "learning_rate": 4.29104226053073e-05, |
| "loss": 0.4566, |
| "step": 80 |
| }, |
| { |
| "epoch": 0.31716417910447764, |
| "grad_norm": 0.7352383281266268, |
| "learning_rate": 4.186775516209732e-05, |
| "loss": 0.4522, |
| "step": 85 |
| }, |
| { |
| "epoch": 0.3358208955223881, |
| "grad_norm": 0.692010393723862, |
| "learning_rate": 4.077015604633669e-05, |
| "loss": 0.4536, |
| "step": 90 |
| }, |
| { |
| "epoch": 0.35447761194029853, |
| "grad_norm": 0.5982685666886067, |
| "learning_rate": 3.962182166550441e-05, |
| "loss": 0.4519, |
| "step": 95 |
| }, |
| { |
| "epoch": 0.373134328358209, |
| "grad_norm": 0.4825080528708406, |
| "learning_rate": 3.8427142401220634e-05, |
| "loss": 0.4488, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.3917910447761194, |
| "grad_norm": 0.6245159548206975, |
| "learning_rate": 3.71906858236735e-05, |
| "loss": 0.4451, |
| "step": 105 |
| }, |
| { |
| "epoch": 0.41044776119402987, |
| "grad_norm": 0.47023483394795307, |
| "learning_rate": 3.591717922860785e-05, |
| "loss": 0.4503, |
| "step": 110 |
| }, |
| { |
| "epoch": 0.4291044776119403, |
| "grad_norm": 0.3489896547326426, |
| "learning_rate": 3.46114915636416e-05, |
| "loss": 0.4425, |
| "step": 115 |
| }, |
| { |
| "epoch": 0.44776119402985076, |
| "grad_norm": 0.4521997610351419, |
| "learning_rate": 3.3278614813010034e-05, |
| "loss": 0.4431, |
| "step": 120 |
| }, |
| { |
| "epoch": 0.4664179104477612, |
| "grad_norm": 0.4377904278160347, |
| "learning_rate": 3.1923644911909e-05, |
| "loss": 0.441, |
| "step": 125 |
| }, |
| { |
| "epoch": 0.48507462686567165, |
| "grad_norm": 0.4114938923473316, |
| "learning_rate": 3.0551762263406576e-05, |
| "loss": 0.4391, |
| "step": 130 |
| }, |
| { |
| "epoch": 0.503731343283582, |
| "grad_norm": 0.4155119032467163, |
| "learning_rate": 2.9168211932412042e-05, |
| "loss": 0.4369, |
| "step": 135 |
| }, |
| { |
| "epoch": 0.5223880597014925, |
| "grad_norm": 0.43137261154438067, |
| "learning_rate": 2.777828359242567e-05, |
| "loss": 0.4409, |
| "step": 140 |
| }, |
| { |
| "epoch": 0.5410447761194029, |
| "grad_norm": 0.33103612314501996, |
| "learning_rate": 2.6387291301738377e-05, |
| "loss": 0.4346, |
| "step": 145 |
| }, |
| { |
| "epoch": 0.5597014925373134, |
| "grad_norm": 0.4314727096833306, |
| "learning_rate": 2.50005531864019e-05, |
| "loss": 0.433, |
| "step": 150 |
| }, |
| { |
| "epoch": 0.5783582089552238, |
| "grad_norm": 0.33914877967482293, |
| "learning_rate": 2.362337110764688e-05, |
| "loss": 0.4325, |
| "step": 155 |
| }, |
| { |
| "epoch": 0.5970149253731343, |
| "grad_norm": 0.3128508821671319, |
| "learning_rate": 2.226101039148557e-05, |
| "loss": 0.4317, |
| "step": 160 |
| }, |
| { |
| "epoch": 0.6156716417910447, |
| "grad_norm": 0.2964801059161522, |
| "learning_rate": 2.0918679697998252e-05, |
| "loss": 0.4301, |
| "step": 165 |
| }, |
| { |
| "epoch": 0.6343283582089553, |
| "grad_norm": 0.34458923690109616, |
| "learning_rate": 1.9601511107268255e-05, |
| "loss": 0.4304, |
| "step": 170 |
| }, |
| { |
| "epoch": 0.6529850746268657, |
| "grad_norm": 0.2952709193193604, |
| "learning_rate": 1.8314540498102216e-05, |
| "loss": 0.4294, |
| "step": 175 |
| }, |
| { |
| "epoch": 0.6716417910447762, |
| "grad_norm": 0.3543986295763423, |
| "learning_rate": 1.7062688294552992e-05, |
| "loss": 0.4236, |
| "step": 180 |
| }, |
| { |
| "epoch": 0.6902985074626866, |
| "grad_norm": 0.3060444359953636, |
| "learning_rate": 1.5850740653856096e-05, |
| "loss": 0.4255, |
| "step": 185 |
| }, |
| { |
| "epoch": 0.7089552238805971, |
| "grad_norm": 0.2825495912702711, |
| "learning_rate": 1.4683331167703218e-05, |
| "loss": 0.4281, |
| "step": 190 |
| }, |
| { |
| "epoch": 0.7276119402985075, |
| "grad_norm": 0.2740450122634198, |
| "learning_rate": 1.356492314681356e-05, |
| "loss": 0.4303, |
| "step": 195 |
| }, |
| { |
| "epoch": 0.746268656716418, |
| "grad_norm": 0.2614619864121355, |
| "learning_rate": 1.2499792556533716e-05, |
| "loss": 0.4253, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.7649253731343284, |
| "grad_norm": 0.2357753416115836, |
| "learning_rate": 1.1492011668707753e-05, |
| "loss": 0.4238, |
| "step": 205 |
| }, |
| { |
| "epoch": 0.7835820895522388, |
| "grad_norm": 0.232641395191967, |
| "learning_rate": 1.0545433492320603e-05, |
| "loss": 0.4273, |
| "step": 210 |
| }, |
| { |
| "epoch": 0.8022388059701493, |
| "grad_norm": 0.21192338187887055, |
| "learning_rate": 9.663677042440537e-06, |
| "loss": 0.4227, |
| "step": 215 |
| }, |
| { |
| "epoch": 0.8208955223880597, |
| "grad_norm": 0.2033015248228483, |
| "learning_rate": 8.850113503781367e-06, |
| "loss": 0.4225, |
| "step": 220 |
| }, |
| { |
| "epoch": 0.8395522388059702, |
| "grad_norm": 0.2011137164412444, |
| "learning_rate": 8.107853341784671e-06, |
| "loss": 0.4274, |
| "step": 225 |
| }, |
| { |
| "epoch": 0.8582089552238806, |
| "grad_norm": 0.25728160053241095, |
| "learning_rate": 7.439734410499752e-06, |
| "loss": 0.4245, |
| "step": 230 |
| }, |
| { |
| "epoch": 0.8768656716417911, |
| "grad_norm": 0.251586337146944, |
| "learning_rate": 6.848311102728011e-06, |
| "loss": 0.4243, |
| "step": 235 |
| }, |
| { |
| "epoch": 0.8955223880597015, |
| "grad_norm": 0.22357913943601826, |
| "learning_rate": 6.335844583913515e-06, |
| "loss": 0.4209, |
| "step": 240 |
| }, |
| { |
| "epoch": 0.914179104477612, |
| "grad_norm": 0.22762126786146625, |
| "learning_rate": 5.904294147118193e-06, |
| "loss": 0.424, |
| "step": 245 |
| }, |
| { |
| "epoch": 0.9328358208955224, |
| "grad_norm": 0.202066770329782, |
| "learning_rate": 5.555309722133842e-06, |
| "loss": 0.421, |
| "step": 250 |
| }, |
| { |
| "epoch": 0.9514925373134329, |
| "grad_norm": 0.19740815078420187, |
| "learning_rate": 5.290225567370509e-06, |
| "loss": 0.4172, |
| "step": 255 |
| }, |
| { |
| "epoch": 0.9701492537313433, |
| "grad_norm": 0.20624100834984413, |
| "learning_rate": 5.110055168638854e-06, |
| "loss": 0.4207, |
| "step": 260 |
| }, |
| { |
| "epoch": 0.9888059701492538, |
| "grad_norm": 0.2003275661172111, |
| "learning_rate": 5.0154873643297575e-06, |
| "loss": 0.4234, |
| "step": 265 |
| }, |
| { |
| "epoch": 1.0, |
| "step": 268, |
| "total_flos": 1.379511412069499e+18, |
| "train_loss": 0.4719447244459124, |
| "train_runtime": 23314.2857, |
| "train_samples_per_second": 1.471, |
| "train_steps_per_second": 0.011 |
| } |
| ], |
| "logging_steps": 5, |
| "max_steps": 268, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 1, |
| "save_steps": 100, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 1.379511412069499e+18, |
| "train_batch_size": 4, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
