{
  "best_metric": 2.4018473625183105,
  "best_model_checkpoint": "./output/training_results/C017_random_sample_Meta-Llama-3-8B_pretrain_20240724/checkpoint-35802",
  "epoch": 4.0,
  "eval_steps": 3978,
  "global_step": 39772,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.00010057326762546515,
      "grad_norm": 0.0,
      "learning_rate": 0.0,
      "loss": 2.7164,
      "step": 1
    },
    {
      "epoch": 0.20004022930705018,
      "grad_norm": 1.9595824499713421,
      "learning_rate": 1.9943010392222595e-06,
      "loss": 2.5584,
      "step": 1989
    },
    {
      "epoch": 0.40008045861410035,
      "grad_norm": 1.8475727372215602,
      "learning_rate": 2.237952084257793e-06,
      "loss": 2.4822,
      "step": 3978
    },
    {
      "epoch": 0.40008045861410035,
      "eval_loss": 2.457235813140869,
      "eval_runtime": 809.2413,
      "eval_samples_per_second": 87.369,
      "eval_steps_per_second": 0.683,
      "step": 3978
    },
    {
      "epoch": 0.6001206879211506,
      "grad_norm": 1.7662911861496735,
      "learning_rate": 1.2175319826671093e-06,
      "loss": 2.4419,
      "step": 5967
    },
    {
      "epoch": 0.8001609172282007,
      "grad_norm": 1.7983814582083166,
      "learning_rate": 6.499668127721108e-07,
      "loss": 2.4215,
      "step": 7956
    },
    {
      "epoch": 0.8001609172282007,
      "eval_loss": 2.4154767990112305,
      "eval_runtime": 808.5614,
      "eval_samples_per_second": 87.443,
      "eval_steps_per_second": 0.684,
      "step": 7956
    },
    {
      "epoch": 1.000201146535251,
      "grad_norm": 1.7969883421186013,
      "learning_rate": 3.4535991044489427e-07,
      "loss": 2.4128,
      "step": 9945
    },
    {
      "epoch": 1.2002413758423012,
      "grad_norm": 1.8600898537077046,
      "learning_rate": 1.8847567083143102e-07,
      "loss": 2.3585,
      "step": 11934
    },
    {
      "epoch": 1.2002413758423012,
      "eval_loss": 2.407348394393921,
      "eval_runtime": 809.2343,
      "eval_samples_per_second": 87.37,
      "eval_steps_per_second": 0.683,
      "step": 11934
    },
    {
      "epoch": 1.4002816051493512,
      "grad_norm": 1.9169506816386161,
      "learning_rate": 1.1133168818945991e-07,
      "loss": 2.3538,
      "step": 13923
    },
    {
      "epoch": 1.6003218344564014,
      "grad_norm": 1.8761137775542636,
      "learning_rate": 7.547427840102147e-08,
      "loss": 2.3526,
      "step": 15912
    },
    {
      "epoch": 1.6003218344564014,
      "eval_loss": 2.40483021736145,
      "eval_runtime": 808.8539,
      "eval_samples_per_second": 87.411,
      "eval_steps_per_second": 0.684,
      "step": 15912
    },
    {
      "epoch": 1.8003620637634516,
      "grad_norm": 1.761004624388482,
      "learning_rate": 5.979347528334647e-08,
      "loss": 2.3525,
      "step": 17901
    },
    {
      "epoch": 2.000402293070502,
      "grad_norm": 1.813766837169378,
      "learning_rate": 5.34373848297726e-08,
      "loss": 2.3542,
      "step": 19890
    },
    {
      "epoch": 2.000402293070502,
      "eval_loss": 2.40364408493042,
      "eval_runtime": 808.719,
      "eval_samples_per_second": 87.426,
      "eval_steps_per_second": 0.684,
      "step": 19890
    },
    {
      "epoch": 2.200442522377552,
      "grad_norm": 1.8622400969038313,
      "learning_rate": 5.108120698150338e-08,
      "loss": 2.3457,
      "step": 21879
    },
    {
      "epoch": 2.4004827516846023,
      "grad_norm": 1.9759538842891853,
      "learning_rate": 5.029702769777328e-08,
      "loss": 2.3441,
      "step": 23868
    },
    {
      "epoch": 2.4004827516846023,
      "eval_loss": 2.4035539627075195,
      "eval_runtime": 809.1929,
      "eval_samples_per_second": 87.375,
      "eval_steps_per_second": 0.683,
      "step": 23868
    },
    {
      "epoch": 2.6005229809916526,
      "grad_norm": 1.9777564770023575,
      "learning_rate": 5.0068636563154646e-08,
      "loss": 2.3446,
      "step": 25857
    },
    {
      "epoch": 2.8005632102987024,
      "grad_norm": 1.8192679060780033,
      "learning_rate": 5.001264640771992e-08,
      "loss": 2.3451,
      "step": 27846
    },
    {
      "epoch": 2.8005632102987024,
      "eval_loss": 2.4028046131134033,
      "eval_runtime": 808.0983,
      "eval_samples_per_second": 87.493,
      "eval_steps_per_second": 0.684,
      "step": 27846
    },
    {
      "epoch": 3.0006034396057526,
      "grad_norm": 2.0137358755507666,
      "learning_rate": 5.0001713989719166e-08,
      "loss": 2.3443,
      "step": 29835
    },
    {
      "epoch": 3.200643668912803,
      "grad_norm": 1.8381002615524122,
      "learning_rate": 5.0000148985750135e-08,
      "loss": 2.3415,
      "step": 31824
    },
    {
      "epoch": 3.200643668912803,
      "eval_loss": 2.402500867843628,
      "eval_runtime": 371.3559,
      "eval_samples_per_second": 190.392,
      "eval_steps_per_second": 1.489,
      "step": 31824
    },
    {
      "epoch": 3.400683898219853,
      "grad_norm": 1.9296099508223659,
      "learning_rate": 5.0000006382565474e-08,
      "loss": 2.34,
      "step": 33813
    },
    {
      "epoch": 3.6007241275269033,
      "grad_norm": 1.9423815954292634,
      "learning_rate": 5.0000000076751155e-08,
      "loss": 2.3406,
      "step": 35802
    },
    {
      "epoch": 3.6007241275269033,
      "eval_loss": 2.4018473625183105,
      "eval_runtime": 346.8193,
      "eval_samples_per_second": 203.861,
      "eval_steps_per_second": 1.594,
      "step": 35802
    },
    {
      "epoch": 3.8007643568339535,
      "grad_norm": 1.8386633630301288,
      "learning_rate": 5.0000000000041456e-08,
      "loss": 2.3406,
      "step": 37791
    },
    {
      "epoch": 4.0,
      "step": 39772,
      "total_flos": 4159012790599680.0,
      "train_loss": 0.7021765150223461,
      "train_runtime": 17437.0559,
      "train_samples_per_second": 145.971,
      "train_steps_per_second": 2.281
    }
  ],
  "logging_steps": 1989,
  "max_steps": 39772,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 3978,
  "total_flos": 4159012790599680.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}