{
  "best_metric": 0.5330753326416016,
  "best_model_checkpoint": "/scratch/project_2006600/dif_models/fi_bert_reg/diff_bert_base_sampled_unfrozen_early_real_test/checkpoint-3500",
  "epoch": 10.666666666666666,
  "eval_steps": 100,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.26666666666666666,
      "grad_norm": 37.87092208862305,
      "learning_rate": 5.444444444444444e-07,
      "loss": 13.2633,
      "step": 100
    },
    {
      "epoch": 0.26666666666666666,
      "eval_loss": 10.746745109558105,
      "eval_mse": 10.746745109558105,
      "eval_r2": -3.070736885070801,
      "eval_rmse": 3.2782227993011475,
      "eval_runtime": 57.8905,
      "eval_samples_per_second": 103.644,
      "eval_steps_per_second": 25.911,
      "step": 100
    },
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 11.965848922729492,
      "learning_rate": 1.0944444444444445e-06,
      "loss": 6.0044,
      "step": 200
    },
    {
      "epoch": 0.5333333333333333,
      "eval_loss": 2.2986323833465576,
      "eval_mse": 2.2986323833465576,
      "eval_r2": 0.129305899143219,
      "eval_rmse": 1.5161241292953491,
      "eval_runtime": 57.8766,
      "eval_samples_per_second": 103.669,
      "eval_steps_per_second": 25.917,
      "step": 200
    },
    {
      "epoch": 0.8,
      "grad_norm": 16.900310516357422,
      "learning_rate": 1.6500000000000003e-06,
      "loss": 1.9777,
      "step": 300
    },
    {
      "epoch": 0.8,
      "eval_loss": 1.0734479427337646,
      "eval_mse": 1.0734479427337646,
      "eval_r2": 0.593390941619873,
      "eval_rmse": 1.0360733270645142,
      "eval_runtime": 57.8679,
      "eval_samples_per_second": 103.684,
      "eval_steps_per_second": 25.921,
      "step": 300
    },
    {
      "epoch": 1.0666666666666667,
      "grad_norm": 36.441951751708984,
      "learning_rate": 2.2e-06,
      "loss": 0.9918,
      "step": 400
    },
    {
      "epoch": 1.0666666666666667,
      "eval_loss": 0.7707585692405701,
      "eval_mse": 0.7707585692405701,
      "eval_r2": 0.708046019077301,
      "eval_rmse": 0.8779285550117493,
      "eval_runtime": 57.8738,
      "eval_samples_per_second": 103.674,
      "eval_steps_per_second": 25.918,
      "step": 400
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 35.2764892578125,
      "learning_rate": 2.755555555555556e-06,
      "loss": 0.7404,
      "step": 500
    },
    {
      "epoch": 1.3333333333333333,
      "eval_loss": 0.8621485829353333,
      "eval_mse": 0.8621485829353333,
      "eval_r2": 0.6734285354614258,
      "eval_rmse": 0.9285195469856262,
      "eval_runtime": 57.8776,
      "eval_samples_per_second": 103.667,
      "eval_steps_per_second": 25.917,
      "step": 500
    },
    {
      "epoch": 1.6,
      "grad_norm": 25.652307510375977,
      "learning_rate": 3.3055555555555558e-06,
      "loss": 0.5897,
      "step": 600
    },
    {
      "epoch": 1.6,
      "eval_loss": 0.7726595997810364,
      "eval_mse": 0.7726595997810364,
      "eval_r2": 0.7073259353637695,
      "eval_rmse": 0.8790105581283569,
      "eval_runtime": 57.8605,
      "eval_samples_per_second": 103.698,
      "eval_steps_per_second": 25.924,
      "step": 600
    },
    {
      "epoch": 1.8666666666666667,
      "grad_norm": 20.405384063720703,
      "learning_rate": 3.861111111111112e-06,
      "loss": 0.5903,
      "step": 700
    },
    {
      "epoch": 1.8666666666666667,
      "eval_loss": 0.7032095789909363,
      "eval_mse": 0.7032095789909363,
      "eval_r2": 0.7336327433586121,
      "eval_rmse": 0.8385758996009827,
      "eval_runtime": 57.8612,
      "eval_samples_per_second": 103.696,
      "eval_steps_per_second": 25.924,
      "step": 700
    },
    {
      "epoch": 2.1333333333333333,
      "grad_norm": 17.7819766998291,
      "learning_rate": 4.416666666666667e-06,
      "loss": 0.4654,
      "step": 800
    },
    {
      "epoch": 2.1333333333333333,
      "eval_loss": 0.6557766199111938,
      "eval_mse": 0.6557766199111938,
      "eval_r2": 0.7515997886657715,
      "eval_rmse": 0.809800386428833,
      "eval_runtime": 57.8596,
      "eval_samples_per_second": 103.699,
      "eval_steps_per_second": 25.925,
      "step": 800
    },
    {
      "epoch": 2.4,
      "grad_norm": 16.64554786682129,
      "learning_rate": 4.9722222222222224e-06,
      "loss": 0.4457,
      "step": 900
    },
    {
      "epoch": 2.4,
      "eval_loss": 0.7730452418327332,
      "eval_mse": 0.7730452418327332,
      "eval_r2": 0.707179844379425,
      "eval_rmse": 0.8792299032211304,
      "eval_runtime": 57.8442,
      "eval_samples_per_second": 103.727,
      "eval_steps_per_second": 25.932,
      "step": 900
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 23.086116790771484,
      "learning_rate": 5.527777777777779e-06,
      "loss": 0.3555,
      "step": 1000
    },
    {
      "epoch": 2.6666666666666665,
      "eval_loss": 0.6905973553657532,
      "eval_mse": 0.6905973553657532,
      "eval_r2": 0.7384101152420044,
      "eval_rmse": 0.8310218453407288,
      "eval_runtime": 57.8673,
      "eval_samples_per_second": 103.685,
      "eval_steps_per_second": 25.921,
      "step": 1000
    },
    {
      "epoch": 2.9333333333333336,
      "grad_norm": 12.183361053466797,
      "learning_rate": 6.083333333333333e-06,
      "loss": 0.3482,
      "step": 1100
    },
    {
      "epoch": 2.9333333333333336,
      "eval_loss": 0.7757130265235901,
      "eval_mse": 0.7757130265235901,
      "eval_r2": 0.7061693072319031,
      "eval_rmse": 0.8807457089424133,
      "eval_runtime": 57.8693,
      "eval_samples_per_second": 103.682,
      "eval_steps_per_second": 25.921,
      "step": 1100
    },
    {
      "epoch": 3.2,
      "grad_norm": 7.989579677581787,
      "learning_rate": 6.6388888888888895e-06,
      "loss": 0.2977,
      "step": 1200
    },
    {
      "epoch": 3.2,
      "eval_loss": 0.756076991558075,
      "eval_mse": 0.7560770511627197,
      "eval_r2": 0.7136071920394897,
      "eval_rmse": 0.8695269227027893,
      "eval_runtime": 57.8756,
      "eval_samples_per_second": 103.671,
      "eval_steps_per_second": 25.918,
      "step": 1200
    },
    {
      "epoch": 3.466666666666667,
      "grad_norm": 23.822872161865234,
      "learning_rate": 7.194444444444445e-06,
      "loss": 0.312,
      "step": 1300
    },
    {
      "epoch": 3.466666666666667,
      "eval_loss": 0.9145962595939636,
      "eval_mse": 0.9145963788032532,
      "eval_r2": 0.6535619497299194,
      "eval_rmse": 0.9563453197479248,
      "eval_runtime": 57.8768,
      "eval_samples_per_second": 103.669,
      "eval_steps_per_second": 25.917,
      "step": 1300
    },
    {
      "epoch": 3.7333333333333334,
      "grad_norm": 15.19206428527832,
      "learning_rate": 7.75e-06,
      "loss": 0.2681,
      "step": 1400
    },
    {
      "epoch": 3.7333333333333334,
      "eval_loss": 0.6494566202163696,
      "eval_mse": 0.6494566202163696,
      "eval_r2": 0.7539936900138855,
      "eval_rmse": 0.8058887124061584,
      "eval_runtime": 57.883,
      "eval_samples_per_second": 103.657,
      "eval_steps_per_second": 25.914,
      "step": 1400
    },
    {
      "epoch": 4.0,
      "grad_norm": 10.404943466186523,
      "learning_rate": 8.305555555555557e-06,
      "loss": 0.2744,
      "step": 1500
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.6918351054191589,
      "eval_mse": 0.6918351054191589,
      "eval_r2": 0.7379412651062012,
      "eval_rmse": 0.8317662477493286,
      "eval_runtime": 57.8842,
      "eval_samples_per_second": 103.655,
      "eval_steps_per_second": 25.914,
      "step": 1500
    },
    {
      "epoch": 4.266666666666667,
      "grad_norm": 11.18205738067627,
      "learning_rate": 8.861111111111111e-06,
      "loss": 0.2142,
      "step": 1600
    },
    {
      "epoch": 4.266666666666667,
      "eval_loss": 0.7790309190750122,
      "eval_mse": 0.779030978679657,
      "eval_r2": 0.7049124836921692,
      "eval_rmse": 0.8826273083686829,
      "eval_runtime": 58.0438,
      "eval_samples_per_second": 103.37,
      "eval_steps_per_second": 25.843,
      "step": 1600
    },
    {
      "epoch": 4.533333333333333,
      "grad_norm": 15.371255874633789,
      "learning_rate": 9.416666666666667e-06,
      "loss": 0.2103,
      "step": 1700
    },
    {
      "epoch": 4.533333333333333,
      "eval_loss": 0.6177367568016052,
      "eval_mse": 0.6177367568016052,
      "eval_r2": 0.7660087943077087,
      "eval_rmse": 0.7859622836112976,
      "eval_runtime": 57.8859,
      "eval_samples_per_second": 103.652,
      "eval_steps_per_second": 25.913,
      "step": 1700
    },
    {
      "epoch": 4.8,
      "grad_norm": 26.033283233642578,
      "learning_rate": 9.972222222222224e-06,
      "loss": 0.2009,
      "step": 1800
    },
    {
      "epoch": 4.8,
      "eval_loss": 0.4925681948661804,
      "eval_mse": 0.4925681948661804,
      "eval_r2": 0.8134211301803589,
      "eval_rmse": 0.7018320560455322,
      "eval_runtime": 57.8886,
      "eval_samples_per_second": 103.647,
      "eval_steps_per_second": 25.912,
      "step": 1800
    },
    {
      "epoch": 5.066666666666666,
      "grad_norm": 14.781371116638184,
      "learning_rate": 1.0527777777777778e-05,
      "loss": 0.2077,
      "step": 1900
    },
    {
      "epoch": 5.066666666666666,
      "eval_loss": 0.731107234954834,
      "eval_mse": 0.7311073541641235,
      "eval_r2": 0.7230653762817383,
      "eval_rmse": 0.8550481796264648,
      "eval_runtime": 57.8849,
      "eval_samples_per_second": 103.654,
      "eval_steps_per_second": 25.913,
      "step": 1900
    },
    {
      "epoch": 5.333333333333333,
      "grad_norm": 10.876953125,
      "learning_rate": 1.1083333333333335e-05,
      "loss": 0.1621,
      "step": 2000
    },
    {
      "epoch": 5.333333333333333,
      "eval_loss": 0.9120460748672485,
      "eval_mse": 0.9120460748672485,
      "eval_r2": 0.654528021812439,
      "eval_rmse": 0.9550110101699829,
      "eval_runtime": 57.8944,
      "eval_samples_per_second": 103.637,
      "eval_steps_per_second": 25.909,
      "step": 2000
    },
    {
      "epoch": 5.6,
      "grad_norm": 9.433772087097168,
      "learning_rate": 1.163888888888889e-05,
      "loss": 0.172,
      "step": 2100
    },
    {
      "epoch": 5.6,
      "eval_loss": 0.7684391140937805,
      "eval_mse": 0.7684392333030701,
      "eval_r2": 0.7089245319366455,
      "eval_rmse": 0.8766066431999207,
      "eval_runtime": 57.8919,
      "eval_samples_per_second": 103.641,
      "eval_steps_per_second": 25.91,
      "step": 2100
    },
    {
      "epoch": 5.866666666666667,
      "grad_norm": 8.749499320983887,
      "learning_rate": 1.2194444444444447e-05,
      "loss": 0.205,
      "step": 2200
    },
    {
      "epoch": 5.866666666666667,
      "eval_loss": 0.9115136861801147,
      "eval_mse": 0.9115136861801147,
      "eval_r2": 0.6547296643257141,
      "eval_rmse": 0.9547322392463684,
      "eval_runtime": 57.8967,
      "eval_samples_per_second": 103.633,
      "eval_steps_per_second": 25.908,
      "step": 2200
    },
    {
      "epoch": 6.133333333333334,
      "grad_norm": 13.432199478149414,
      "learning_rate": 1.275e-05,
      "loss": 0.1548,
      "step": 2300
    },
    {
      "epoch": 6.133333333333334,
      "eval_loss": 0.7451738119125366,
      "eval_mse": 0.7451738119125366,
      "eval_r2": 0.7177371978759766,
      "eval_rmse": 0.8632345199584961,
      "eval_runtime": 57.8927,
      "eval_samples_per_second": 103.64,
      "eval_steps_per_second": 25.91,
      "step": 2300
    },
    {
      "epoch": 6.4,
      "grad_norm": 13.872139930725098,
      "learning_rate": 1.3305555555555556e-05,
      "loss": 0.1593,
      "step": 2400
    },
    {
      "epoch": 6.4,
      "eval_loss": 0.6607070565223694,
      "eval_mse": 0.6607070565223694,
      "eval_r2": 0.7497321963310242,
      "eval_rmse": 0.8128389120101929,
      "eval_runtime": 57.8803,
      "eval_samples_per_second": 103.662,
      "eval_steps_per_second": 25.916,
      "step": 2400
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 10.801892280578613,
      "learning_rate": 1.3855555555555558e-05,
      "loss": 0.1684,
      "step": 2500
    },
    {
      "epoch": 6.666666666666667,
      "eval_loss": 0.8161285519599915,
      "eval_mse": 0.8161285519599915,
      "eval_r2": 0.690860390663147,
      "eval_rmse": 0.903398334980011,
      "eval_runtime": 57.8953,
      "eval_samples_per_second": 103.635,
      "eval_steps_per_second": 25.909,
      "step": 2500
    },
    {
      "epoch": 6.933333333333334,
      "grad_norm": 13.156838417053223,
      "learning_rate": 1.441111111111111e-05,
      "loss": 0.1823,
      "step": 2600
    },
    {
      "epoch": 6.933333333333334,
      "eval_loss": 1.0073966979980469,
      "eval_mse": 1.0073966979980469,
      "eval_r2": 0.6184103488922119,
      "eval_rmse": 1.003691554069519,
      "eval_runtime": 57.9026,
      "eval_samples_per_second": 103.622,
      "eval_steps_per_second": 25.906,
      "step": 2600
    },
    {
      "epoch": 7.2,
      "grad_norm": 16.002506256103516,
      "learning_rate": 1.4966666666666667e-05,
      "loss": 0.1936,
      "step": 2700
    },
    {
      "epoch": 7.2,
      "eval_loss": 0.8159008622169495,
      "eval_mse": 0.8159008622169495,
      "eval_r2": 0.690946638584137,
      "eval_rmse": 0.9032723307609558,
      "eval_runtime": 57.9032,
      "eval_samples_per_second": 103.621,
      "eval_steps_per_second": 25.905,
      "step": 2700
    },
    {
      "epoch": 7.466666666666667,
      "grad_norm": 12.294295310974121,
      "learning_rate": 1.5522222222222223e-05,
      "loss": 0.1506,
      "step": 2800
    },
    {
      "epoch": 7.466666666666667,
      "eval_loss": 0.8324363827705383,
      "eval_mse": 0.8324363827705383,
      "eval_r2": 0.6846832036972046,
      "eval_rmse": 0.9123795032501221,
      "eval_runtime": 57.9057,
      "eval_samples_per_second": 103.617,
      "eval_steps_per_second": 25.904,
      "step": 2800
    },
    {
      "epoch": 7.733333333333333,
      "grad_norm": 5.195356369018555,
      "learning_rate": 1.607777777777778e-05,
      "loss": 0.1499,
      "step": 2900
    },
    {
      "epoch": 7.733333333333333,
      "eval_loss": 0.8613407611846924,
      "eval_mse": 0.8613407611846924,
      "eval_r2": 0.6737345457077026,
      "eval_rmse": 0.9280844330787659,
      "eval_runtime": 57.9123,
      "eval_samples_per_second": 103.605,
      "eval_steps_per_second": 25.901,
      "step": 2900
    },
    {
      "epoch": 8.0,
      "grad_norm": 11.357749938964844,
      "learning_rate": 1.6633333333333336e-05,
      "loss": 0.1631,
      "step": 3000
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.6052080988883972,
      "eval_mse": 0.6052080988883972,
      "eval_r2": 0.7707545161247253,
      "eval_rmse": 0.7779512405395508,
      "eval_runtime": 57.8963,
      "eval_samples_per_second": 103.634,
      "eval_steps_per_second": 25.908,
      "step": 3000
    },
    {
      "epoch": 8.266666666666667,
      "grad_norm": 20.62454605102539,
      "learning_rate": 1.7188888888888892e-05,
      "loss": 0.1559,
      "step": 3100
    },
    {
      "epoch": 8.266666666666667,
      "eval_loss": 0.584992527961731,
      "eval_mse": 0.584992527961731,
      "eval_r2": 0.7784119248390198,
      "eval_rmse": 0.7648480534553528,
      "eval_runtime": 57.8939,
      "eval_samples_per_second": 103.638,
      "eval_steps_per_second": 25.909,
      "step": 3100
    },
    {
      "epoch": 8.533333333333333,
      "grad_norm": 5.493505954742432,
      "learning_rate": 1.7744444444444445e-05,
      "loss": 0.1434,
      "step": 3200
    },
    {
      "epoch": 8.533333333333333,
      "eval_loss": 0.6216039061546326,
      "eval_mse": 0.6216039061546326,
      "eval_r2": 0.7645439505577087,
      "eval_rmse": 0.7884185910224915,
      "eval_runtime": 57.9149,
      "eval_samples_per_second": 103.6,
      "eval_steps_per_second": 25.9,
      "step": 3200
    },
    {
      "epoch": 8.8,
      "grad_norm": 8.887247085571289,
      "learning_rate": 1.83e-05,
      "loss": 0.1485,
      "step": 3300
    },
    {
      "epoch": 8.8,
      "eval_loss": 0.6688755750656128,
      "eval_mse": 0.6688756346702576,
      "eval_r2": 0.7466380000114441,
      "eval_rmse": 0.8178481459617615,
      "eval_runtime": 57.8949,
      "eval_samples_per_second": 103.636,
      "eval_steps_per_second": 25.909,
      "step": 3300
    },
    {
      "epoch": 9.066666666666666,
      "grad_norm": 54.376461029052734,
      "learning_rate": 1.8855555555555557e-05,
      "loss": 0.1487,
      "step": 3400
    },
    {
      "epoch": 9.066666666666666,
      "eval_loss": 0.6742914915084839,
      "eval_mse": 0.6742914915084839,
      "eval_r2": 0.7445865869522095,
      "eval_rmse": 0.8211525678634644,
      "eval_runtime": 57.9237,
      "eval_samples_per_second": 103.584,
      "eval_steps_per_second": 25.896,
      "step": 3400
    },
    {
      "epoch": 9.333333333333334,
      "grad_norm": 2.6493732929229736,
      "learning_rate": 1.941111111111111e-05,
      "loss": 0.1248,
      "step": 3500
    },
    {
      "epoch": 9.333333333333334,
      "eval_loss": 0.5330753326416016,
      "eval_mse": 0.5330753326416016,
      "eval_r2": 0.7980775237083435,
      "eval_rmse": 0.730120062828064,
      "eval_runtime": 57.9187,
      "eval_samples_per_second": 103.593,
      "eval_steps_per_second": 25.898,
      "step": 3500
    },
    {
      "epoch": 9.6,
      "grad_norm": 22.65760040283203,
      "learning_rate": 1.9966666666666666e-05,
      "loss": 0.127,
      "step": 3600
    },
    {
      "epoch": 9.6,
      "eval_loss": 0.760678231716156,
      "eval_mse": 0.760678231716156,
      "eval_r2": 0.7118643522262573,
      "eval_rmse": 0.8721687197685242,
      "eval_runtime": 57.9121,
      "eval_samples_per_second": 103.605,
      "eval_steps_per_second": 25.901,
      "step": 3600
    },
    {
      "epoch": 9.866666666666667,
      "grad_norm": 11.861601829528809,
      "learning_rate": 1.999958463257905e-05,
      "loss": 0.1409,
      "step": 3700
    },
    {
      "epoch": 9.866666666666667,
      "eval_loss": 0.6627658009529114,
      "eval_mse": 0.6627658009529114,
      "eval_r2": 0.748952329158783,
      "eval_rmse": 0.8141043186187744,
      "eval_runtime": 57.9148,
      "eval_samples_per_second": 103.601,
      "eval_steps_per_second": 25.9,
      "step": 3700
    },
    {
      "epoch": 10.133333333333333,
      "grad_norm": 10.758604049682617,
      "learning_rate": 1.999823082667316e-05,
      "loss": 0.1352,
      "step": 3800
    },
    {
      "epoch": 10.133333333333333,
      "eval_loss": 0.7061165571212769,
      "eval_mse": 0.7061165571212769,
      "eval_r2": 0.7325316071510315,
      "eval_rmse": 0.8403074145317078,
      "eval_runtime": 58.8761,
      "eval_samples_per_second": 101.909,
      "eval_steps_per_second": 25.477,
      "step": 3800
    },
    {
      "epoch": 10.4,
      "grad_norm": 4.267393112182617,
      "learning_rate": 1.999593701724414e-05,
      "loss": 0.1418,
      "step": 3900
    },
    {
      "epoch": 10.4,
      "eval_loss": 0.6572955846786499,
      "eval_mse": 0.6572955846786499,
      "eval_r2": 0.7510244250297546,
      "eval_rmse": 0.810737669467926,
      "eval_runtime": 57.921,
      "eval_samples_per_second": 103.589,
      "eval_steps_per_second": 25.897,
      "step": 3900
    },
    {
      "epoch": 10.666666666666666,
      "grad_norm": 4.308215618133545,
      "learning_rate": 1.9992703419949032e-05,
      "loss": 0.138,
      "step": 4000
    },
    {
      "epoch": 10.666666666666666,
      "eval_loss": 0.6316142678260803,
      "eval_mse": 0.6316142678260803,
      "eval_r2": 0.7607522010803223,
      "eval_rmse": 0.7947416305541992,
      "eval_runtime": 57.92,
      "eval_samples_per_second": 103.591,
      "eval_steps_per_second": 25.898,
      "step": 4000
    }
  ],
  "logging_steps": 100,
  "max_steps": 9000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 24,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 6,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.6858183041024e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}