{
  "best_metric": 0.5,
  "best_model_checkpoint": "./zephyr/10-04-24-Weni-WeniGPT-Agents-Zephyr-1.0.25-KTO_Experiment with a new tokenizer configuration for chat template of zephyr-2_max_steps-1470_batch_16_2024-04-10_ppid_9/checkpoint-300",
  "epoch": 5.2805280528052805,
  "eval_steps": 50,
  "global_step": 800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.13,
      "grad_norm": 57.293792724609375,
      "kl": 0.03853478282690048,
      "learning_rate": 6.222222222222222e-05,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7078,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 20
    },
    {
      "epoch": 0.26,
      "grad_norm": 112.50944519042969,
      "kl": 3.2648494243621826,
      "learning_rate": 0.00014666666666666666,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6966,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 40
    },
    {
      "epoch": 0.33,
      "eval_kl": 0.0,
      "eval_logps/chosen": -413.6161193847656,
      "eval_logps/rejected": -362.2559509277344,
      "eval_loss": 0.5063381791114807,
      "eval_rewards/chosen": -13.412939071655273,
      "eval_rewards/margins": -1.0048810243606567,
      "eval_rewards/rejected": -12.408059120178223,
      "eval_runtime": 170.1826,
      "eval_samples_per_second": 2.057,
      "eval_steps_per_second": 0.517,
      "step": 50
    },
    {
      "epoch": 0.4,
      "grad_norm": 19.94582748413086,
      "kl": 0.45922356843948364,
      "learning_rate": 0.00019887719298245616,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.5743,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 60
    },
    {
      "epoch": 0.53,
      "grad_norm": 79.92957305908203,
      "kl": 0.0,
      "learning_rate": 0.0001960701754385965,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6108,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 80
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.06103940308094025,
      "kl": 0.0,
      "learning_rate": 0.00019326315789473686,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.754,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 100
    },
    {
      "epoch": 0.66,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2027.0018310546875,
      "eval_logps/rejected": -1697.82177734375,
      "eval_loss": 0.5000000596046448,
      "eval_rewards/chosen": -174.75149536132812,
      "eval_rewards/margins": -28.786863327026367,
      "eval_rewards/rejected": -145.96463012695312,
      "eval_runtime": 170.0562,
      "eval_samples_per_second": 2.058,
      "eval_steps_per_second": 0.517,
      "step": 100
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.0001904561403508772,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.95,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 120
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00018764912280701756,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6274,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 140
    },
    {
      "epoch": 0.99,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2237.81494140625,
      "eval_logps/rejected": -1889.774169921875,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -195.83285522460938,
      "eval_rewards/margins": -30.672954559326172,
      "eval_rewards/rejected": -165.15989685058594,
      "eval_runtime": 169.8795,
      "eval_samples_per_second": 2.06,
      "eval_steps_per_second": 0.518,
      "step": 150
    },
    {
      "epoch": 1.06,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.0001848421052631579,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6387,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 160
    },
    {
      "epoch": 1.19,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00018203508771929826,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.8327,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 180
    },
    {
      "epoch": 1.32,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00017922807017543862,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.642,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 200
    },
    {
      "epoch": 1.32,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2230.916259765625,
      "eval_logps/rejected": -1884.9520263671875,
      "eval_loss": 0.5000000596046448,
      "eval_rewards/chosen": -195.14297485351562,
      "eval_rewards/margins": -30.465293884277344,
      "eval_rewards/rejected": -164.67767333984375,
      "eval_runtime": 170.1489,
      "eval_samples_per_second": 2.057,
      "eval_steps_per_second": 0.517,
      "step": 200
    },
    {
      "epoch": 1.45,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00017642105263157896,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7493,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 220
    },
    {
      "epoch": 1.58,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.0001736140350877193,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6241,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 240
    },
    {
      "epoch": 1.65,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2230.957275390625,
      "eval_logps/rejected": -1885.0225830078125,
      "eval_loss": 0.5000000596046448,
      "eval_rewards/chosen": -195.14706420898438,
      "eval_rewards/margins": -30.462318420410156,
      "eval_rewards/rejected": -164.68475341796875,
      "eval_runtime": 170.1092,
      "eval_samples_per_second": 2.058,
      "eval_steps_per_second": 0.517,
      "step": 250
    },
    {
      "epoch": 1.72,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00017080701754385965,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.9621,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 260
    },
    {
      "epoch": 1.85,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.000168,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7279,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 280
    },
    {
      "epoch": 1.98,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00016519298245614035,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7477,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 300
    },
    {
      "epoch": 1.98,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2238.164306640625,
      "eval_logps/rejected": -1890.7996826171875,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -195.86773681640625,
      "eval_rewards/margins": -30.605329513549805,
      "eval_rewards/rejected": -165.26242065429688,
      "eval_runtime": 170.0647,
      "eval_samples_per_second": 2.058,
      "eval_steps_per_second": 0.517,
      "step": 300
    },
    {
      "epoch": 2.11,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00016238596491228072,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7111,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 320
    },
    {
      "epoch": 2.24,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00015957894736842105,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.8685,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 340
    },
    {
      "epoch": 2.31,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2238.054931640625,
      "eval_logps/rejected": -1890.694580078125,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -195.85682678222656,
      "eval_rewards/margins": -30.604921340942383,
      "eval_rewards/rejected": -165.2519073486328,
      "eval_runtime": 170.0528,
      "eval_samples_per_second": 2.058,
      "eval_steps_per_second": 0.517,
      "step": 350
    },
    {
      "epoch": 2.38,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00015677192982456142,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6905,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 360
    },
    {
      "epoch": 2.51,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00015396491228070175,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.736,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 380
    },
    {
      "epoch": 2.64,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00015115789473684211,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.693,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 400
    },
    {
      "epoch": 2.64,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2237.827392578125,
      "eval_logps/rejected": -1890.5028076171875,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -195.83407592773438,
      "eval_rewards/margins": -30.601318359375,
      "eval_rewards/rejected": -165.23275756835938,
      "eval_runtime": 170.2445,
      "eval_samples_per_second": 2.056,
      "eval_steps_per_second": 0.517,
      "step": 400
    },
    {
      "epoch": 2.77,
      "grad_norm": 8.788210266175156e-07,
      "kl": 0.0,
      "learning_rate": 0.00014835087719298245,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.8652,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 420
    },
    {
      "epoch": 2.9,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.0001455438596491228,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.686,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 440
    },
    {
      "epoch": 2.97,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2237.722412109375,
      "eval_logps/rejected": -1890.4027099609375,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -195.82354736328125,
      "eval_rewards/margins": -30.600812911987305,
      "eval_rewards/rejected": -165.22274780273438,
      "eval_runtime": 170.3429,
      "eval_samples_per_second": 2.055,
      "eval_steps_per_second": 0.517,
      "step": 450
    },
    {
      "epoch": 3.04,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00014273684210526318,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6858,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 460
    },
    {
      "epoch": 3.17,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.0001399298245614035,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.8479,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 480
    },
    {
      "epoch": 3.3,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00013712280701754388,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6119,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 500
    },
    {
      "epoch": 3.3,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2237.6083984375,
      "eval_logps/rejected": -1890.3140869140625,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -195.81216430664062,
      "eval_rewards/margins": -30.598268508911133,
      "eval_rewards/rejected": -165.21388244628906,
      "eval_runtime": 169.9488,
      "eval_samples_per_second": 2.059,
      "eval_steps_per_second": 0.518,
      "step": 500
    },
    {
      "epoch": 3.43,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.0001343157894736842,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7107,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 520
    },
    {
      "epoch": 3.56,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00013150877192982455,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.5902,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 540
    },
    {
      "epoch": 3.63,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2237.56494140625,
      "eval_logps/rejected": -1890.3043212890625,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -195.80784606933594,
      "eval_rewards/margins": -30.59491539001465,
      "eval_rewards/rejected": -165.21290588378906,
      "eval_runtime": 169.9756,
      "eval_samples_per_second": 2.059,
      "eval_steps_per_second": 0.518,
      "step": 550
    },
    {
      "epoch": 3.7,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.0001287017543859649,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.9042,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 560
    },
    {
      "epoch": 3.83,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00012589473684210527,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7268,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 580
    },
    {
      "epoch": 3.96,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00012308771929824564,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7106,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 600
    },
    {
      "epoch": 3.96,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2241.97509765625,
      "eval_logps/rejected": -1893.87646484375,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -196.24884033203125,
      "eval_rewards/margins": -30.67871856689453,
      "eval_rewards/rejected": -165.57012939453125,
      "eval_runtime": 169.9427,
      "eval_samples_per_second": 2.06,
      "eval_steps_per_second": 0.518,
      "step": 600
    },
    {
      "epoch": 4.09,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00012028070175438597,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6829,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 620
    },
    {
      "epoch": 4.22,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00011747368421052631,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.8232,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 640
    },
    {
      "epoch": 4.29,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2241.91552734375,
      "eval_logps/rejected": -1893.757080078125,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -196.24290466308594,
      "eval_rewards/margins": -30.684709548950195,
      "eval_rewards/rejected": -165.55816650390625,
      "eval_runtime": 169.9605,
      "eval_samples_per_second": 2.059,
      "eval_steps_per_second": 0.518,
      "step": 650
    },
    {
      "epoch": 4.36,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00011466666666666667,
      "logps/chosen": -2123.240234375,
      "logps/rejected": NaN,
      "loss": 0.6315,
      "rewards/chosen": -188.09486389160156,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 660
    },
    {
      "epoch": 4.49,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00011185964912280702,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7998,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 680
    },
    {
      "epoch": 4.62,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00010905263157894738,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.5881,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 700
    },
    {
      "epoch": 4.62,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2251.134033203125,
      "eval_logps/rejected": -1901.2047119140625,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -197.1647491455078,
      "eval_rewards/margins": -30.8618221282959,
      "eval_rewards/rejected": -166.30291748046875,
      "eval_runtime": 169.9704,
      "eval_samples_per_second": 2.059,
      "eval_steps_per_second": 0.518,
      "step": 700
    },
    {
      "epoch": 4.75,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00010624561403508772,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.8756,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 720
    },
    {
      "epoch": 4.88,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00010343859649122807,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6156,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 740
    },
    {
      "epoch": 4.95,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2250.90234375,
      "eval_logps/rejected": -1901.0179443359375,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -197.1415557861328,
      "eval_rewards/margins": -30.857322692871094,
      "eval_rewards/rejected": -166.28424072265625,
      "eval_runtime": 169.9616,
      "eval_samples_per_second": 2.059,
      "eval_steps_per_second": 0.518,
      "step": 750
    },
    {
      "epoch": 5.02,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 0.00010063157894736843,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7376,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 760
    },
    {
      "epoch": 5.15,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 9.782456140350877e-05,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.7998,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 780
    },
    {
      "epoch": 5.28,
      "grad_norm": 0.0,
      "kl": 0.0,
      "learning_rate": 9.501754385964913e-05,
      "logps/chosen": NaN,
      "logps/rejected": NaN,
      "loss": 0.6291,
      "rewards/chosen": NaN,
      "rewards/margins": NaN,
      "rewards/rejected": NaN,
      "step": 800
    },
    {
      "epoch": 5.28,
      "eval_kl": 0.0,
      "eval_logps/chosen": -2250.995849609375,
      "eval_logps/rejected": -1901.1036376953125,
      "eval_loss": 0.5,
      "eval_rewards/chosen": -197.15087890625,
      "eval_rewards/margins": -30.858049392700195,
      "eval_rewards/rejected": -166.29283142089844,
      "eval_runtime": 169.941,
      "eval_samples_per_second": 2.06,
      "eval_steps_per_second": 0.518,
      "step": 800
    }
  ],
  "logging_steps": 20,
  "max_steps": 1470,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}