{
  "best_metric": 0.8500627352572145,
  "best_model_checkpoint": "./hatebert_finetuned_v5/checkpoint-15000",
  "epoch": 1.8799348289259306,
  "eval_steps": 1000,
  "global_step": 15000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06266449429753101,
      "grad_norm": 2.7858245372772217,
      "learning_rate": 4.160401002506266e-06,
      "loss": 0.5658,
      "step": 500
    },
    {
      "epoch": 0.12532898859506203,
      "grad_norm": 0.6836428642272949,
      "learning_rate": 8.320802005012531e-06,
      "loss": 0.3394,
      "step": 1000
    },
    {
      "epoch": 0.12532898859506203,
      "eval_accuracy": 0.9509008303305656,
      "eval_f1": 0.7696604439217992,
      "eval_loss": 0.3321492373943329,
      "eval_optimal_threshold": 0.7856788039207458,
      "eval_runtime": 24.5664,
      "eval_samples_per_second": 1299.13,
      "eval_steps_per_second": 40.625,
      "step": 1000
    },
    {
      "epoch": 0.18799348289259304,
      "grad_norm": 0.07630700618028641,
      "learning_rate": 1.2497911445279867e-05,
      "loss": 0.3098,
      "step": 1500
    },
    {
      "epoch": 0.25065797719012406,
      "grad_norm": 0.6884164214134216,
      "learning_rate": 1.66750208855472e-05,
      "loss": 0.3245,
      "step": 2000
    },
    {
      "epoch": 0.25065797719012406,
      "eval_accuracy": 0.9620241265862447,
      "eval_f1": 0.8181272509003601,
      "eval_loss": 0.24252600967884064,
      "eval_optimal_threshold": 0.6846858263015747,
      "eval_runtime": 24.5401,
      "eval_samples_per_second": 1300.526,
      "eval_steps_per_second": 40.668,
      "step": 2000
    },
    {
      "epoch": 0.3133224714876551,
      "grad_norm": 1.6269655227661133,
      "learning_rate": 1.9906234043540826e-05,
      "loss": 0.2939,
      "step": 2500
    },
    {
      "epoch": 0.3759869657851861,
      "grad_norm": 53.418663024902344,
      "learning_rate": 1.9442046140277588e-05,
      "loss": 0.2911,
      "step": 3000
    },
    {
      "epoch": 0.3759869657851861,
      "eval_accuracy": 0.9669434435218549,
      "eval_f1": 0.8339889850511408,
      "eval_loss": 0.32369741797447205,
      "eval_optimal_threshold": 0.639145016670227,
      "eval_runtime": 24.5157,
      "eval_samples_per_second": 1301.819,
      "eval_steps_per_second": 40.709,
      "step": 3000
    },
    {
      "epoch": 0.43865146008271716,
      "grad_norm": 2.552913188934326,
      "learning_rate": 1.8977858237014345e-05,
      "loss": 0.2717,
      "step": 3500
    },
    {
      "epoch": 0.5013159543802481,
      "grad_norm": 74.53817749023438,
      "learning_rate": 1.8513670333751103e-05,
      "loss": 0.2574,
      "step": 4000
    },
    {
      "epoch": 0.5013159543802481,
      "eval_accuracy": 0.9656587811373962,
      "eval_f1": 0.8311768330252619,
      "eval_loss": 0.344524085521698,
      "eval_optimal_threshold": 0.903037965297699,
      "eval_runtime": 24.5127,
      "eval_samples_per_second": 1301.981,
      "eval_steps_per_second": 40.714,
      "step": 4000
    },
    {
      "epoch": 0.5639804486777792,
      "grad_norm": 4.5688018798828125,
      "learning_rate": 1.8049482430487864e-05,
      "loss": 0.2455,
      "step": 4500
    },
    {
      "epoch": 0.6266449429753101,
      "grad_norm": 0.18218272924423218,
      "learning_rate": 1.7585294527224622e-05,
      "loss": 0.278,
      "step": 5000
    },
    {
      "epoch": 0.6266449429753101,
      "eval_accuracy": 0.966818110606298,
      "eval_f1": 0.8378005820186859,
      "eval_loss": 0.22682912647724152,
      "eval_optimal_threshold": 0.8996956944465637,
      "eval_runtime": 24.5217,
      "eval_samples_per_second": 1301.5,
      "eval_steps_per_second": 40.699,
      "step": 5000
    },
    {
      "epoch": 0.6893094372728412,
      "grad_norm": 1.167153239250183,
      "learning_rate": 1.712110662396138e-05,
      "loss": 0.2859,
      "step": 5500
    },
    {
      "epoch": 0.7519739315703722,
      "grad_norm": 8.231012344360352,
      "learning_rate": 1.665691872069814e-05,
      "loss": 0.2335,
      "step": 6000
    },
    {
      "epoch": 0.7519739315703722,
      "eval_accuracy": 0.9660661131129563,
      "eval_f1": 0.8360333080999243,
      "eval_loss": 0.2758038341999054,
      "eval_optimal_threshold": 0.3707093894481659,
      "eval_runtime": 24.5444,
      "eval_samples_per_second": 1300.296,
      "eval_steps_per_second": 40.661,
      "step": 6000
    },
    {
      "epoch": 0.8146384258679032,
      "grad_norm": 8.103411674499512,
      "learning_rate": 1.6193659193241424e-05,
      "loss": 0.2617,
      "step": 6500
    },
    {
      "epoch": 0.8773029201654343,
      "grad_norm": 0.008309279568493366,
      "learning_rate": 1.5729471289978185e-05,
      "loss": 0.2656,
      "step": 7000
    },
    {
      "epoch": 0.8773029201654343,
      "eval_accuracy": 0.966034779884067,
      "eval_f1": 0.8370904718966036,
      "eval_loss": 0.28418678045272827,
      "eval_optimal_threshold": 0.9179669618606567,
      "eval_runtime": 24.512,
      "eval_samples_per_second": 1302.016,
      "eval_steps_per_second": 40.715,
      "step": 7000
    },
    {
      "epoch": 0.9399674144629653,
      "grad_norm": 11.311033248901367,
      "learning_rate": 1.5265283386714943e-05,
      "loss": 0.2464,
      "step": 7500
    },
    {
      "epoch": 1.0026319087604962,
      "grad_norm": 0.2538510262966156,
      "learning_rate": 1.4801095483451701e-05,
      "loss": 0.2474,
      "step": 8000
    },
    {
      "epoch": 1.0026319087604962,
      "eval_accuracy": 0.9657841140529532,
      "eval_f1": 0.8362327534493101,
      "eval_loss": 0.29536861181259155,
      "eval_optimal_threshold": 0.9757624268531799,
      "eval_runtime": 24.5401,
      "eval_samples_per_second": 1300.525,
      "eval_steps_per_second": 40.668,
      "step": 8000
    },
    {
      "epoch": 1.0652964030580274,
      "grad_norm": 0.09177704900503159,
      "learning_rate": 1.4336907580188462e-05,
      "loss": 0.1691,
      "step": 8500
    },
    {
      "epoch": 1.1279608973555584,
      "grad_norm": 0.3952280580997467,
      "learning_rate": 1.3872719676925222e-05,
      "loss": 0.1862,
      "step": 9000
    },
    {
      "epoch": 1.1279608973555584,
      "eval_accuracy": 0.9682907723640921,
      "eval_f1": 0.843440594059406,
      "eval_loss": 0.298546701669693,
      "eval_optimal_threshold": 0.8941442370414734,
      "eval_runtime": 24.557,
      "eval_samples_per_second": 1299.63,
      "eval_steps_per_second": 40.64,
      "step": 9000
    },
    {
      "epoch": 1.1906253916530893,
      "grad_norm": 0.009977204725146294,
      "learning_rate": 1.340853177366198e-05,
      "loss": 0.1579,
      "step": 9500
    },
    {
      "epoch": 1.2532898859506203,
      "grad_norm": 0.2561907470226288,
      "learning_rate": 1.2944343870398739e-05,
      "loss": 0.1869,
      "step": 10000
    },
    {
      "epoch": 1.2532898859506203,
      "eval_accuracy": 0.9682594391352028,
      "eval_f1": 0.8454143140546315,
      "eval_loss": 0.3232070207595825,
      "eval_optimal_threshold": 0.9672868847846985,
      "eval_runtime": 24.5414,
      "eval_samples_per_second": 1300.454,
      "eval_steps_per_second": 40.666,
      "step": 10000
    },
    {
      "epoch": 1.3159543802481515,
      "grad_norm": 0.14401866495609283,
      "learning_rate": 1.2480155967135497e-05,
      "loss": 0.1901,
      "step": 10500
    },
    {
      "epoch": 1.3786188745456824,
      "grad_norm": 0.20385725796222687,
      "learning_rate": 1.2015968063872256e-05,
      "loss": 0.173,
      "step": 11000
    },
    {
      "epoch": 1.3786188745456824,
      "eval_accuracy": 0.9673194422685257,
      "eval_f1": 0.8431343059106633,
      "eval_loss": 0.26260536909103394,
      "eval_optimal_threshold": 0.9728307723999023,
      "eval_runtime": 24.51,
      "eval_samples_per_second": 1302.121,
      "eval_steps_per_second": 40.718,
      "step": 11000
    },
    {
      "epoch": 1.4412833688432134,
      "grad_norm": 0.26596102118492126,
      "learning_rate": 1.1552708536415542e-05,
      "loss": 0.1654,
      "step": 11500
    },
    {
      "epoch": 1.5039478631407444,
      "grad_norm": 0.09112328290939331,
      "learning_rate": 1.10885206331523e-05,
      "loss": 0.2088,
      "step": 12000
    },
    {
      "epoch": 1.5039478631407444,
      "eval_accuracy": 0.9675387748707505,
      "eval_f1": 0.8438347904733192,
      "eval_loss": 0.30822300910949707,
      "eval_optimal_threshold": 0.9813465476036072,
      "eval_runtime": 24.5129,
      "eval_samples_per_second": 1301.97,
      "eval_steps_per_second": 40.713,
      "step": 12000
    },
    {
      "epoch": 1.5666123574382755,
      "grad_norm": 0.1223333552479744,
      "learning_rate": 1.0625261105695586e-05,
      "loss": 0.168,
      "step": 12500
    },
    {
      "epoch": 1.6292768517358065,
      "grad_norm": 21.354354858398438,
      "learning_rate": 1.0161073202432346e-05,
      "loss": 0.1621,
      "step": 13000
    },
    {
      "epoch": 1.6292768517358065,
      "eval_accuracy": 0.9618361272129093,
      "eval_f1": 0.82609937178755,
      "eval_loss": 0.2638513147830963,
      "eval_optimal_threshold": 0.9734938144683838,
      "eval_runtime": 24.4962,
      "eval_samples_per_second": 1302.854,
      "eval_steps_per_second": 40.741,
      "step": 13000
    },
    {
      "epoch": 1.6919413460333375,
      "grad_norm": 0.11861222982406616,
      "learning_rate": 9.696885299169104e-06,
      "loss": 0.1757,
      "step": 13500
    },
    {
      "epoch": 1.7546058403308686,
      "grad_norm": 0.12898346781730652,
      "learning_rate": 9.232697395905863e-06,
      "loss": 0.1665,
      "step": 14000
    },
    {
      "epoch": 1.7546058403308686,
      "eval_accuracy": 0.9669434435218549,
      "eval_f1": 0.8423726281189302,
      "eval_loss": 0.3221518099308014,
      "eval_optimal_threshold": 0.92950439453125,
      "eval_runtime": 24.5035,
      "eval_samples_per_second": 1302.465,
      "eval_steps_per_second": 40.729,
      "step": 14000
    },
    {
      "epoch": 1.8172703346283996,
      "grad_norm": 0.4160967171192169,
      "learning_rate": 8.76943786844915e-06,
      "loss": 0.185,
      "step": 14500
    },
    {
      "epoch": 1.8799348289259306,
      "grad_norm": 0.013835388235747814,
      "learning_rate": 8.305249965185907e-06,
      "loss": 0.1556,
      "step": 15000
    },
    {
      "epoch": 1.8799348289259306,
      "eval_accuracy": 0.9700454331818894,
      "eval_f1": 0.8500627352572145,
      "eval_loss": 0.3804091215133667,
      "eval_optimal_threshold": 0.7121821045875549,
      "eval_runtime": 24.4818,
      "eval_samples_per_second": 1303.622,
      "eval_steps_per_second": 40.765,
      "step": 15000
    }
  ],
  "logging_steps": 500,
  "max_steps": 23937,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 1000,
  "total_flos": 0.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}