{
  "best_metric": 0.6332046332046332,
  "best_model_checkpoint": "./Validated_cracks_22122025_outputs/checkpoint-1620",
  "epoch": 30.0,
  "eval_steps": 500,
  "global_step": 3240,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.46296296296296297,
      "grad_norm": 10.873927116394043,
      "learning_rate": 2.901234567901235e-06,
      "loss": 1.415,
      "step": 50
    },
    {
      "epoch": 0.9259259259259259,
      "grad_norm": 10.678427696228027,
      "learning_rate": 5.9876543209876546e-06,
      "loss": 1.3724,
      "step": 100
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.3359073359073359,
      "eval_loss": 1.3528215885162354,
      "eval_runtime": 1.7714,
      "eval_samples_per_second": 146.211,
      "eval_steps_per_second": 18.629,
      "step": 108
    },
    {
      "epoch": 1.3888888888888888,
      "grad_norm": 15.727510452270508,
      "learning_rate": 9.074074074074075e-06,
      "loss": 1.3621,
      "step": 150
    },
    {
      "epoch": 1.8518518518518519,
      "grad_norm": 18.13157844543457,
      "learning_rate": 9.996809987196146e-06,
      "loss": 1.3528,
      "step": 200
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.42084942084942084,
      "eval_loss": 1.2848700284957886,
      "eval_runtime": 1.764,
      "eval_samples_per_second": 146.821,
      "eval_steps_per_second": 18.707,
      "step": 216
    },
    {
      "epoch": 2.314814814814815,
      "grad_norm": 11.158724784851074,
      "learning_rate": 9.981195232868493e-06,
      "loss": 1.2858,
      "step": 250
    },
    {
      "epoch": 2.7777777777777777,
      "grad_norm": 27.68938446044922,
      "learning_rate": 9.952610423187516e-06,
      "loss": 1.2839,
      "step": 300
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.46332046332046334,
      "eval_loss": 1.2019163370132446,
      "eval_runtime": 1.7428,
      "eval_samples_per_second": 148.61,
      "eval_steps_per_second": 18.935,
      "step": 324
    },
    {
      "epoch": 3.240740740740741,
      "grad_norm": 16.059484481811523,
      "learning_rate": 9.91208532978737e-06,
      "loss": 1.2499,
      "step": 350
    },
    {
      "epoch": 3.7037037037037037,
      "grad_norm": 10.013123512268066,
      "learning_rate": 9.85807175279907e-06,
      "loss": 1.2608,
      "step": 400
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5212355212355212,
      "eval_loss": 1.1559516191482544,
      "eval_runtime": 1.7474,
      "eval_samples_per_second": 148.218,
      "eval_steps_per_second": 18.885,
      "step": 432
    },
    {
      "epoch": 4.166666666666667,
      "grad_norm": 24.753084182739258,
      "learning_rate": 9.791408709849578e-06,
      "loss": 1.2172,
      "step": 450
    },
    {
      "epoch": 4.62962962962963,
      "grad_norm": 12.23658275604248,
      "learning_rate": 9.712269778427157e-06,
      "loss": 1.2178,
      "step": 500
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.555984555984556,
      "eval_loss": 1.0907317399978638,
      "eval_runtime": 1.747,
      "eval_samples_per_second": 148.252,
      "eval_steps_per_second": 18.889,
      "step": 540
    },
    {
      "epoch": 5.092592592592593,
      "grad_norm": 15.365643501281738,
      "learning_rate": 9.620861020786583e-06,
      "loss": 1.2052,
      "step": 550
    },
    {
      "epoch": 5.555555555555555,
      "grad_norm": 12.787246704101562,
      "learning_rate": 9.517420447403444e-06,
      "loss": 1.1624,
      "step": 600
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.528957528957529,
      "eval_loss": 1.0542538166046143,
      "eval_runtime": 1.7382,
      "eval_samples_per_second": 149.005,
      "eval_steps_per_second": 18.985,
      "step": 648
    },
    {
      "epoch": 6.018518518518518,
      "grad_norm": 13.462636947631836,
      "learning_rate": 9.404634840109069e-06,
      "loss": 1.166,
      "step": 650
    },
    {
      "epoch": 6.481481481481482,
      "grad_norm": 15.236122131347656,
      "learning_rate": 9.278195395448948e-06,
      "loss": 1.1335,
      "step": 700
    },
    {
      "epoch": 6.944444444444445,
      "grad_norm": 19.687719345092773,
      "learning_rate": 9.140616369122732e-06,
      "loss": 1.1108,
      "step": 750
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.5598455598455598,
      "eval_loss": 1.0452436208724976,
      "eval_runtime": 1.7531,
      "eval_samples_per_second": 147.737,
      "eval_steps_per_second": 18.824,
      "step": 756
    },
    {
      "epoch": 7.407407407407407,
      "grad_norm": 16.47871971130371,
      "learning_rate": 8.992255989929632e-06,
      "loss": 1.0708,
      "step": 800
    },
    {
      "epoch": 7.87037037037037,
      "grad_norm": 19.686279296875,
      "learning_rate": 8.833500559197024e-06,
      "loss": 1.1028,
      "step": 850
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.5598455598455598,
      "eval_loss": 1.0568522214889526,
      "eval_runtime": 1.7528,
      "eval_samples_per_second": 147.764,
      "eval_steps_per_second": 18.827,
      "step": 864
    },
    {
      "epoch": 8.333333333333334,
      "grad_norm": 37.79273986816406,
      "learning_rate": 8.664763444927562e-06,
      "loss": 1.0619,
      "step": 900
    },
    {
      "epoch": 8.796296296296296,
      "grad_norm": 16.420129776000977,
      "learning_rate": 8.486484005469977e-06,
      "loss": 1.1023,
      "step": 950
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.5675675675675675,
      "eval_loss": 1.057991862297058,
      "eval_runtime": 1.7353,
      "eval_samples_per_second": 149.25,
      "eval_steps_per_second": 19.016,
      "step": 972
    },
    {
      "epoch": 9.25925925925926,
      "grad_norm": 25.430509567260742,
      "learning_rate": 8.299126445516126e-06,
      "loss": 1.1072,
      "step": 1000
    },
    {
      "epoch": 9.722222222222221,
      "grad_norm": 35.4229850769043,
      "learning_rate": 8.103178607403005e-06,
      "loss": 1.0572,
      "step": 1050
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.6138996138996139,
      "eval_loss": 1.0031245946884155,
      "eval_runtime": 1.743,
      "eval_samples_per_second": 148.597,
      "eval_steps_per_second": 18.933,
      "step": 1080
    },
    {
      "epoch": 10.185185185185185,
      "grad_norm": 14.466697692871094,
      "learning_rate": 7.899150700867014e-06,
      "loss": 1.0489,
      "step": 1100
    },
    {
      "epoch": 10.648148148148149,
      "grad_norm": 21.01825714111328,
      "learning_rate": 7.687573974557857e-06,
      "loss": 0.9874,
      "step": 1150
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.583011583011583,
      "eval_loss": 1.0340404510498047,
      "eval_runtime": 1.7486,
      "eval_samples_per_second": 148.117,
      "eval_steps_per_second": 18.872,
      "step": 1188
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 16.48180389404297,
      "learning_rate": 7.4689993327712765e-06,
      "loss": 1.0212,
      "step": 1200
    },
    {
      "epoch": 11.574074074074074,
      "grad_norm": 18.663524627685547,
      "learning_rate": 7.243995901002312e-06,
      "loss": 1.0132,
      "step": 1250
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.61003861003861,
      "eval_loss": 1.004916787147522,
      "eval_runtime": 1.7428,
      "eval_samples_per_second": 148.608,
      "eval_steps_per_second": 18.935,
      "step": 1296
    },
    {
      "epoch": 12.037037037037036,
      "grad_norm": 24.026588439941406,
      "learning_rate": 7.013149544054148e-06,
      "loss": 1.0111,
      "step": 1300
    },
    {
      "epoch": 12.5,
      "grad_norm": 20.475515365600586,
      "learning_rate": 6.777061340561082e-06,
      "loss": 1.005,
      "step": 1350
    },
    {
      "epoch": 12.962962962962964,
      "grad_norm": 49.801326751708984,
      "learning_rate": 6.5363460178976524e-06,
      "loss": 0.9898,
      "step": 1400
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.6216216216216216,
      "eval_loss": 0.9874952435493469,
      "eval_runtime": 1.7492,
      "eval_samples_per_second": 148.065,
      "eval_steps_per_second": 18.865,
      "step": 1404
    },
    {
      "epoch": 13.425925925925926,
      "grad_norm": 38.209232330322266,
      "learning_rate": 6.291630351549136e-06,
      "loss": 0.9332,
      "step": 1450
    },
    {
      "epoch": 13.88888888888889,
      "grad_norm": 16.34885025024414,
      "learning_rate": 6.043551533111121e-06,
      "loss": 1.0182,
      "step": 1500
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.61003861003861,
      "eval_loss": 0.9667864441871643,
      "eval_runtime": 1.7502,
      "eval_samples_per_second": 147.984,
      "eval_steps_per_second": 18.855,
      "step": 1512
    },
    {
      "epoch": 14.351851851851851,
      "grad_norm": 21.03826904296875,
      "learning_rate": 5.792755511167572e-06,
      "loss": 0.9279,
      "step": 1550
    },
    {
      "epoch": 14.814814814814815,
      "grad_norm": 34.984375,
      "learning_rate": 5.544968491028696e-06,
      "loss": 0.9889,
      "step": 1600
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.6332046332046332,
      "eval_loss": 0.9692270755767822,
      "eval_runtime": 1.7309,
      "eval_samples_per_second": 149.63,
      "eval_steps_per_second": 19.065,
      "step": 1620
    },
    {
      "epoch": 15.277777777777779,
      "grad_norm": 16.456146240234375,
      "learning_rate": 5.290724144552379e-06,
      "loss": 0.9673,
      "step": 1650
    },
    {
      "epoch": 15.74074074074074,
      "grad_norm": 27.009721755981445,
      "learning_rate": 5.035722809427533e-06,
      "loss": 0.9446,
      "step": 1700
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.6332046332046332,
      "eval_loss": 0.9777135848999023,
      "eval_runtime": 1.7383,
      "eval_samples_per_second": 148.995,
      "eval_steps_per_second": 18.984,
      "step": 1728
    },
    {
      "epoch": 16.203703703703702,
      "grad_norm": 12.319013595581055,
      "learning_rate": 4.780628459113764e-06,
      "loss": 0.9024,
      "step": 1750
    },
    {
      "epoch": 16.666666666666668,
      "grad_norm": 20.682802200317383,
      "learning_rate": 4.526105309263983e-06,
      "loss": 0.9519,
      "step": 1800
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.5984555984555985,
      "eval_loss": 1.0030242204666138,
      "eval_runtime": 1.7421,
      "eval_samples_per_second": 148.671,
      "eval_steps_per_second": 18.943,
      "step": 1836
    },
    {
      "epoch": 17.12962962962963,
      "grad_norm": 24.291671752929688,
      "learning_rate": 4.272816088237135e-06,
      "loss": 1.003,
      "step": 1850
    },
    {
      "epoch": 17.59259259259259,
      "grad_norm": 13.384744644165039,
      "learning_rate": 4.021420311483538e-06,
      "loss": 0.9458,
      "step": 1900
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.5984555984555985,
      "eval_loss": 0.97477126121521,
      "eval_runtime": 1.7465,
      "eval_samples_per_second": 148.299,
      "eval_steps_per_second": 18.895,
      "step": 1944
    },
    {
      "epoch": 18.055555555555557,
      "grad_norm": 15.239733695983887,
      "learning_rate": 3.7725725642960047e-06,
      "loss": 0.9798,
      "step": 1950
    },
    {
      "epoch": 18.51851851851852,
      "grad_norm": 24.09659767150879,
      "learning_rate": 3.526920797398148e-06,
      "loss": 0.9346,
      "step": 2000
    },
    {
      "epoch": 18.98148148148148,
      "grad_norm": 23.59765625,
      "learning_rate": 3.2851046398077705e-06,
      "loss": 0.9347,
      "step": 2050
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.6177606177606177,
      "eval_loss": 0.9744410514831543,
      "eval_runtime": 1.7456,
      "eval_samples_per_second": 148.373,
      "eval_steps_per_second": 18.905,
      "step": 2052
    },
    {
      "epoch": 19.444444444444443,
      "grad_norm": 22.252914428710938,
      "learning_rate": 3.0477537333683815e-06,
      "loss": 0.8767,
      "step": 2100
    },
    {
      "epoch": 19.90740740740741,
      "grad_norm": 32.37730026245117,
      "learning_rate": 2.815486093285317e-06,
      "loss": 0.8863,
      "step": 2150
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.6293436293436293,
      "eval_loss": 0.9656945466995239,
      "eval_runtime": 1.7376,
      "eval_samples_per_second": 149.058,
      "eval_steps_per_second": 18.992,
      "step": 2160
    },
    {
      "epoch": 20.37037037037037,
      "grad_norm": 18.95844841003418,
      "learning_rate": 2.5889064989353797e-06,
      "loss": 0.9081,
      "step": 2200
    },
    {
      "epoch": 20.833333333333332,
      "grad_norm": 84.69548034667969,
      "learning_rate": 2.3686049191399614e-06,
      "loss": 0.8507,
      "step": 2250
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.6254826254826255,
      "eval_loss": 0.9784243702888489,
      "eval_runtime": 1.7492,
      "eval_samples_per_second": 148.065,
      "eval_steps_per_second": 18.865,
      "step": 2268
    },
    {
      "epoch": 21.296296296296298,
      "grad_norm": 34.8240852355957,
      "learning_rate": 2.155154976001948e-06,
      "loss": 0.9308,
      "step": 2300
    },
    {
      "epoch": 21.75925925925926,
      "grad_norm": 39.87207794189453,
      "learning_rate": 1.949112451306282e-06,
      "loss": 0.8712,
      "step": 2350
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.6254826254826255,
      "eval_loss": 0.9790379405021667,
      "eval_runtime": 1.7458,
      "eval_samples_per_second": 148.355,
      "eval_steps_per_second": 18.902,
      "step": 2376
    },
    {
      "epoch": 22.22222222222222,
      "grad_norm": 15.800605773925781,
      "learning_rate": 1.7510138393732029e-06,
      "loss": 0.8973,
      "step": 2400
    },
    {
      "epoch": 22.685185185185187,
      "grad_norm": 13.362822532653809,
      "learning_rate": 1.5613749501322705e-06,
      "loss": 0.8857,
      "step": 2450
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.6177606177606177,
      "eval_loss": 0.9682068824768066,
      "eval_runtime": 1.7485,
      "eval_samples_per_second": 148.127,
      "eval_steps_per_second": 18.873,
      "step": 2484
    },
    {
      "epoch": 23.14814814814815,
      "grad_norm": 26.928611755371094,
      "learning_rate": 1.3806895660544805e-06,
      "loss": 0.8446,
      "step": 2500
    },
    {
      "epoch": 23.61111111111111,
      "grad_norm": 52.092063903808594,
      "learning_rate": 1.2094281564395628e-06,
      "loss": 0.8848,
      "step": 2550
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.6254826254826255,
      "eval_loss": 0.9722690582275391,
      "eval_runtime": 1.7534,
      "eval_samples_per_second": 147.714,
      "eval_steps_per_second": 18.821,
      "step": 2592
    },
    {
      "epoch": 24.074074074074073,
      "grad_norm": 22.29905128479004,
      "learning_rate": 1.0480366524062041e-06,
      "loss": 0.8778,
      "step": 2600
    },
    {
      "epoch": 24.537037037037038,
      "grad_norm": 23.949947357177734,
      "learning_rate": 8.969352857748842e-07,
      "loss": 0.8672,
      "step": 2650
    },
    {
      "epoch": 25.0,
      "grad_norm": 31.42759895324707,
      "learning_rate": 7.565174948666382e-07,
      "loss": 0.904,
      "step": 2700
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.6177606177606177,
      "eval_loss": 0.975362241268158,
      "eval_runtime": 1.7452,
      "eval_samples_per_second": 148.407,
      "eval_steps_per_second": 18.909,
      "step": 2700
    },
    {
      "epoch": 25.462962962962962,
      "grad_norm": 20.711061477661133,
      "learning_rate": 6.271489000668418e-07,
      "loss": 0.7932,
      "step": 2750
    },
    {
      "epoch": 25.925925925925927,
      "grad_norm": 22.23466682434082,
      "learning_rate": 5.091663518214407e-07,
      "loss": 0.943,
      "step": 2800
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.6216216216216216,
      "eval_loss": 0.9710479974746704,
      "eval_runtime": 1.7435,
      "eval_samples_per_second": 148.554,
      "eval_steps_per_second": 18.928,
      "step": 2808
    },
    {
      "epoch": 26.38888888888889,
      "grad_norm": 18.16496467590332,
      "learning_rate": 4.0287705354446147e-07,
      "loss": 0.9108,
      "step": 2850
    },
    {
      "epoch": 26.85185185185185,
      "grad_norm": 50.363128662109375,
      "learning_rate": 3.085577617205765e-07,
      "loss": 0.862,
      "step": 2900
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.6177606177606177,
      "eval_loss": 0.9716824889183044,
      "eval_runtime": 1.7298,
      "eval_samples_per_second": 149.728,
      "eval_steps_per_second": 19.077,
      "step": 2916
    },
    {
      "epoch": 27.314814814814813,
      "grad_norm": 69.7729721069336,
      "learning_rate": 2.2645406528550407e-07,
      "loss": 0.8412,
      "step": 2950
    },
    {
      "epoch": 27.77777777777778,
      "grad_norm": 20.13194465637207,
      "learning_rate": 1.5677974616058856e-07,
      "loss": 0.864,
      "step": 3000
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.6177606177606177,
      "eval_loss": 0.9705125689506531,
      "eval_runtime": 1.7418,
      "eval_samples_per_second": 148.699,
      "eval_steps_per_second": 18.946,
      "step": 3024
    },
    {
      "epoch": 28.24074074074074,
      "grad_norm": 29.41867446899414,
      "learning_rate": 9.971622260661007e-08,
      "loss": 0.8656,
      "step": 3050
    },
    {
      "epoch": 28.703703703703702,
      "grad_norm": 19.62738609313965,
      "learning_rate": 5.541207684621908e-08,
      "loss": 0.8879,
      "step": 3100
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.6177606177606177,
      "eval_loss": 0.97031569480896,
      "eval_runtime": 1.7373,
      "eval_samples_per_second": 149.084,
      "eval_steps_per_second": 18.995,
      "step": 3132
    },
    {
      "epoch": 29.166666666666668,
      "grad_norm": 18.350698471069336,
      "learning_rate": 2.398266818496864e-08,
      "loss": 0.8594,
      "step": 3150
    },
    {
      "epoch": 29.62962962962963,
      "grad_norm": 28.054824829101562,
      "learning_rate": 5.509832638314061e-09,
      "loss": 0.8099,
      "step": 3200
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.6177606177606177,
      "eval_loss": 0.97026127576828,
      "eval_runtime": 1.8767,
      "eval_samples_per_second": 138.009,
      "eval_steps_per_second": 17.584,
      "step": 3240
    },
    {
      "epoch": 30.0,
      "step": 3240,
      "total_flos": 2.0405460235862016e+18,
      "train_loss": 1.012899116233543,
      "train_runtime": 640.9857,
      "train_samples_per_second": 40.251,
      "train_steps_per_second": 5.055
    }
  ],
  "logging_steps": 50,
  "max_steps": 3240,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 30,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 2.0405460235862016e+18,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}