{
  "best_metric": 0.12666618824005127,
  "best_model_checkpoint": "idbwtiny/checkpoint-20100",
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 20100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.25,
      "grad_norm": 2.5227627754211426,
      "learning_rate": 0.00019503482587064676,
      "loss": 0.4184,
      "step": 500
    },
    {
      "epoch": 0.5,
      "grad_norm": 2.693358898162842,
      "learning_rate": 0.0001900597014925373,
      "loss": 0.4215,
      "step": 1000
    },
    {
      "epoch": 0.75,
      "grad_norm": 1.985125184059143,
      "learning_rate": 0.00018508457711442785,
      "loss": 0.3987,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "grad_norm": 1.946892261505127,
      "learning_rate": 0.0001801094527363184,
      "loss": 0.3871,
      "step": 2000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8674150542486978,
      "eval_loss": 0.3020988404750824,
      "eval_runtime": 206.2751,
      "eval_samples_per_second": 257.816,
      "eval_steps_per_second": 32.229,
      "step": 2010
    },
    {
      "epoch": 1.24,
      "grad_norm": 2.351902723312378,
      "learning_rate": 0.00017514427860696518,
      "loss": 0.3685,
      "step": 2500
    },
    {
      "epoch": 1.49,
      "grad_norm": 2.192362070083618,
      "learning_rate": 0.00017016915422885572,
      "loss": 0.3615,
      "step": 3000
    },
    {
      "epoch": 1.74,
      "grad_norm": 2.392155647277832,
      "learning_rate": 0.00016519402985074627,
      "loss": 0.3474,
      "step": 3500
    },
    {
      "epoch": 1.99,
      "grad_norm": 2.2455360889434814,
      "learning_rate": 0.00016021890547263681,
      "loss": 0.3403,
      "step": 4000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8966548203305692,
      "eval_loss": 0.24423933029174805,
      "eval_runtime": 204.2994,
      "eval_samples_per_second": 260.309,
      "eval_steps_per_second": 32.54,
      "step": 4020
    },
    {
      "epoch": 2.24,
      "grad_norm": 2.0239338874816895,
      "learning_rate": 0.0001552437810945274,
      "loss": 0.3266,
      "step": 4500
    },
    {
      "epoch": 2.49,
      "grad_norm": 2.220857858657837,
      "learning_rate": 0.00015027860696517414,
      "loss": 0.316,
      "step": 5000
    },
    {
      "epoch": 2.74,
      "grad_norm": 1.665713906288147,
      "learning_rate": 0.00014530348258706468,
      "loss": 0.3089,
      "step": 5500
    },
    {
      "epoch": 2.99,
      "grad_norm": 2.8135392665863037,
      "learning_rate": 0.00014032835820895523,
      "loss": 0.3058,
      "step": 6000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9052105075120814,
      "eval_loss": 0.2269268035888672,
      "eval_runtime": 209.4743,
      "eval_samples_per_second": 253.878,
      "eval_steps_per_second": 31.737,
      "step": 6030
    },
    {
      "epoch": 3.23,
      "grad_norm": 2.630223512649536,
      "learning_rate": 0.000135363184079602,
      "loss": 0.2868,
      "step": 6500
    },
    {
      "epoch": 3.48,
      "grad_norm": 1.8844317197799683,
      "learning_rate": 0.00013038805970149256,
      "loss": 0.2843,
      "step": 7000
    },
    {
      "epoch": 3.73,
      "grad_norm": 2.541440010070801,
      "learning_rate": 0.0001254129353233831,
      "loss": 0.2794,
      "step": 7500
    },
    {
      "epoch": 3.98,
      "grad_norm": 2.5618529319763184,
      "learning_rate": 0.00012043781094527365,
      "loss": 0.2722,
      "step": 8000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9196893627423328,
      "eval_loss": 0.20079384744167328,
      "eval_runtime": 208.4042,
      "eval_samples_per_second": 255.182,
      "eval_steps_per_second": 31.9,
      "step": 8040
    },
    {
      "epoch": 4.23,
      "grad_norm": 1.9443061351776123,
      "learning_rate": 0.0001154726368159204,
      "loss": 0.2546,
      "step": 8500
    },
    {
      "epoch": 4.48,
      "grad_norm": 1.7675588130950928,
      "learning_rate": 0.00011049751243781094,
      "loss": 0.2498,
      "step": 9000
    },
    {
      "epoch": 4.73,
      "grad_norm": 2.6675548553466797,
      "learning_rate": 0.00010552238805970149,
      "loss": 0.2472,
      "step": 9500
    },
    {
      "epoch": 4.98,
      "grad_norm": 3.0125935077667236,
      "learning_rate": 0.00010054726368159203,
      "loss": 0.2496,
      "step": 10000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9271920422707358,
      "eval_loss": 0.184106707572937,
      "eval_runtime": 208.5204,
      "eval_samples_per_second": 255.04,
      "eval_steps_per_second": 31.882,
      "step": 10050
    },
    {
      "epoch": 5.22,
      "grad_norm": 3.092007637023926,
      "learning_rate": 9.55721393034826e-05,
      "loss": 0.2271,
      "step": 10500
    },
    {
      "epoch": 5.47,
      "grad_norm": 4.371829509735107,
      "learning_rate": 9.059701492537314e-05,
      "loss": 0.2264,
      "step": 11000
    },
    {
      "epoch": 5.72,
      "grad_norm": 2.334900140762329,
      "learning_rate": 8.562189054726368e-05,
      "loss": 0.2257,
      "step": 11500
    },
    {
      "epoch": 5.97,
      "grad_norm": 3.219524621963501,
      "learning_rate": 8.064676616915423e-05,
      "loss": 0.2219,
      "step": 12000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9352588330418758,
      "eval_loss": 0.1706165224313736,
      "eval_runtime": 212.161,
      "eval_samples_per_second": 250.663,
      "eval_steps_per_second": 31.335,
      "step": 12060
    },
    {
      "epoch": 6.22,
      "grad_norm": 4.32515287399292,
      "learning_rate": 7.5681592039801e-05,
      "loss": 0.207,
      "step": 12500
    },
    {
      "epoch": 6.47,
      "grad_norm": 2.8301429748535156,
      "learning_rate": 7.070646766169154e-05,
      "loss": 0.1996,
      "step": 13000
    },
    {
      "epoch": 6.72,
      "grad_norm": 4.194028377532959,
      "learning_rate": 6.573134328358209e-05,
      "loss": 0.2018,
      "step": 13500
    },
    {
      "epoch": 6.97,
      "grad_norm": 3.551224708557129,
      "learning_rate": 6.075621890547264e-05,
      "loss": 0.198,
      "step": 14000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9451871909140482,
      "eval_loss": 0.147793248295784,
      "eval_runtime": 210.5297,
      "eval_samples_per_second": 252.606,
      "eval_steps_per_second": 31.577,
      "step": 14070
    },
    {
      "epoch": 7.21,
      "grad_norm": 2.972999095916748,
      "learning_rate": 5.5791044776119405e-05,
      "loss": 0.1892,
      "step": 14500
    },
    {
      "epoch": 7.46,
      "grad_norm": 2.5483310222625732,
      "learning_rate": 5.081592039800995e-05,
      "loss": 0.1845,
      "step": 15000
    },
    {
      "epoch": 7.71,
      "grad_norm": 2.5494730472564697,
      "learning_rate": 4.58407960199005e-05,
      "loss": 0.1848,
      "step": 15500
    },
    {
      "epoch": 7.96,
      "grad_norm": 4.215138912200928,
      "learning_rate": 4.086567164179105e-05,
      "loss": 0.1812,
      "step": 16000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9503206032229555,
      "eval_loss": 0.1412646621465683,
      "eval_runtime": 211.1854,
      "eval_samples_per_second": 251.821,
      "eval_steps_per_second": 31.479,
      "step": 16080
    },
    {
      "epoch": 8.21,
      "grad_norm": 3.0078041553497314,
      "learning_rate": 3.5900497512437814e-05,
      "loss": 0.1719,
      "step": 16500
    },
    {
      "epoch": 8.46,
      "grad_norm": 3.556603193283081,
      "learning_rate": 3.092537313432836e-05,
      "loss": 0.1717,
      "step": 17000
    },
    {
      "epoch": 8.71,
      "grad_norm": 2.1547062397003174,
      "learning_rate": 2.595024875621891e-05,
      "loss": 0.1682,
      "step": 17500
    },
    {
      "epoch": 8.96,
      "grad_norm": 2.7211575508117676,
      "learning_rate": 2.0975124378109454e-05,
      "loss": 0.1632,
      "step": 18000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.955303585867133,
      "eval_loss": 0.13010455667972565,
      "eval_runtime": 201.7962,
      "eval_samples_per_second": 263.538,
      "eval_steps_per_second": 32.944,
      "step": 18090
    },
    {
      "epoch": 9.2,
      "grad_norm": 2.98066782951355,
      "learning_rate": 1.600995024875622e-05,
      "loss": 0.1626,
      "step": 18500
    },
    {
      "epoch": 9.45,
      "grad_norm": 2.36365008354187,
      "learning_rate": 1.1034825870646767e-05,
      "loss": 0.1599,
      "step": 19000
    },
    {
      "epoch": 9.7,
      "grad_norm": 3.5116114616394043,
      "learning_rate": 6.059701492537314e-06,
      "loss": 0.1568,
      "step": 19500
    },
    {
      "epoch": 9.95,
      "grad_norm": 2.070329427719116,
      "learning_rate": 1.0845771144278609e-06,
      "loss": 0.1576,
      "step": 20000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9568830973467968,
      "eval_loss": 0.12666618824005127,
      "eval_runtime": 201.0618,
      "eval_samples_per_second": 264.501,
      "eval_steps_per_second": 33.064,
      "step": 20100
    },
    {
      "epoch": 10.0,
      "step": 20100,
      "total_flos": 7.572439992003492e+19,
      "train_loss": 0.2540779383265557,
      "train_runtime": 33188.7721,
      "train_samples_per_second": 90.801,
      "train_steps_per_second": 0.606
    }
  ],
  "logging_steps": 500,
  "max_steps": 20100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 7.572439992003492e+19,
  "train_batch_size": 150,
  "trial_name": null,
  "trial_params": null
}