{
  "best_metric": 0.40972957015037537,
  "best_model_checkpoint": "mikhail-panzo/zlm_b32_le5_s4000/checkpoint-4000",
  "epoch": 1.675392670157068,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.020942408376963352,
      "grad_norm": 5.243492126464844,
      "learning_rate": 2.4000000000000003e-07,
      "loss": 1.1863,
      "step": 50
    },
    {
      "epoch": 0.041884816753926704,
      "grad_norm": 16.012319564819336,
      "learning_rate": 4.900000000000001e-07,
      "loss": 1.01,
      "step": 100
    },
    {
      "epoch": 0.06282722513089005,
      "grad_norm": 3.0757930278778076,
      "learning_rate": 7.4e-07,
      "loss": 0.9741,
      "step": 150
    },
    {
      "epoch": 0.08376963350785341,
      "grad_norm": 3.897144079208374,
      "learning_rate": 9.85e-07,
      "loss": 0.9358,
      "step": 200
    },
    {
      "epoch": 0.10471204188481675,
      "grad_norm": 3.6317360401153564,
      "learning_rate": 1.235e-06,
      "loss": 0.8767,
      "step": 250
    },
    {
      "epoch": 0.1256544502617801,
      "grad_norm": 4.750243186950684,
      "learning_rate": 1.485e-06,
      "loss": 0.8576,
      "step": 300
    },
    {
      "epoch": 0.14659685863874344,
      "grad_norm": 3.4898314476013184,
      "learning_rate": 1.7350000000000001e-06,
      "loss": 0.8472,
      "step": 350
    },
    {
      "epoch": 0.16753926701570682,
      "grad_norm": 4.961293697357178,
      "learning_rate": 1.985e-06,
      "loss": 0.7591,
      "step": 400
    },
    {
      "epoch": 0.18848167539267016,
      "grad_norm": 3.254897117614746,
      "learning_rate": 2.235e-06,
      "loss": 0.7548,
      "step": 450
    },
    {
      "epoch": 0.2094240837696335,
      "grad_norm": 3.3995440006256104,
      "learning_rate": 2.4850000000000003e-06,
      "loss": 0.7211,
      "step": 500
    },
    {
      "epoch": 0.2094240837696335,
      "eval_loss": 0.6148238182067871,
      "eval_runtime": 267.7517,
      "eval_samples_per_second": 31.705,
      "eval_steps_per_second": 3.966,
      "step": 500
    },
    {
      "epoch": 0.23036649214659685,
      "grad_norm": 3.2476744651794434,
      "learning_rate": 2.7350000000000005e-06,
      "loss": 0.6859,
      "step": 550
    },
    {
      "epoch": 0.2513089005235602,
      "grad_norm": 4.222187042236328,
      "learning_rate": 2.9850000000000002e-06,
      "loss": 0.6653,
      "step": 600
    },
    {
      "epoch": 0.27225130890052357,
      "grad_norm": 3.4279122352600098,
      "learning_rate": 3.2350000000000004e-06,
      "loss": 0.6612,
      "step": 650
    },
    {
      "epoch": 0.2931937172774869,
      "grad_norm": 2.7588794231414795,
      "learning_rate": 3.485e-06,
      "loss": 0.6404,
      "step": 700
    },
    {
      "epoch": 0.31413612565445026,
      "grad_norm": 3.481127977371216,
      "learning_rate": 3.7350000000000002e-06,
      "loss": 0.647,
      "step": 750
    },
    {
      "epoch": 0.33507853403141363,
      "grad_norm": 4.153278827667236,
      "learning_rate": 3.985000000000001e-06,
      "loss": 0.6374,
      "step": 800
    },
    {
      "epoch": 0.35602094240837695,
      "grad_norm": 5.10654354095459,
      "learning_rate": 4.235e-06,
      "loss": 0.6225,
      "step": 850
    },
    {
      "epoch": 0.3769633507853403,
      "grad_norm": 2.7014458179473877,
      "learning_rate": 4.485e-06,
      "loss": 0.6026,
      "step": 900
    },
    {
      "epoch": 0.39790575916230364,
      "grad_norm": 2.965581178665161,
      "learning_rate": 4.735e-06,
      "loss": 0.5963,
      "step": 950
    },
    {
      "epoch": 0.418848167539267,
      "grad_norm": 2.671508312225342,
      "learning_rate": 4.9850000000000006e-06,
      "loss": 0.6059,
      "step": 1000
    },
    {
      "epoch": 0.418848167539267,
      "eval_loss": 0.514046847820282,
      "eval_runtime": 268.9563,
      "eval_samples_per_second": 31.563,
      "eval_steps_per_second": 3.949,
      "step": 1000
    },
    {
      "epoch": 0.4397905759162304,
      "grad_norm": 2.840752363204956,
      "learning_rate": 5.235e-06,
      "loss": 0.5759,
      "step": 1050
    },
    {
      "epoch": 0.4607329842931937,
      "grad_norm": 4.1088714599609375,
      "learning_rate": 5.485e-06,
      "loss": 0.5678,
      "step": 1100
    },
    {
      "epoch": 0.4816753926701571,
      "grad_norm": 2.7752063274383545,
      "learning_rate": 5.735e-06,
      "loss": 0.5759,
      "step": 1150
    },
    {
      "epoch": 0.5026178010471204,
      "grad_norm": 5.213489055633545,
      "learning_rate": 5.985000000000001e-06,
      "loss": 0.5476,
      "step": 1200
    },
    {
      "epoch": 0.5235602094240838,
      "grad_norm": 2.6381824016571045,
      "learning_rate": 6.235000000000001e-06,
      "loss": 0.5632,
      "step": 1250
    },
    {
      "epoch": 0.5445026178010471,
      "grad_norm": 3.088671922683716,
      "learning_rate": 6.485000000000001e-06,
      "loss": 0.5625,
      "step": 1300
    },
    {
      "epoch": 0.5654450261780105,
      "grad_norm": 3.4818124771118164,
      "learning_rate": 6.735000000000001e-06,
      "loss": 0.5608,
      "step": 1350
    },
    {
      "epoch": 0.5863874345549738,
      "grad_norm": 3.6994926929473877,
      "learning_rate": 6.985000000000001e-06,
      "loss": 0.5351,
      "step": 1400
    },
    {
      "epoch": 0.6073298429319371,
      "grad_norm": 4.822711944580078,
      "learning_rate": 7.235000000000001e-06,
      "loss": 0.5407,
      "step": 1450
    },
    {
      "epoch": 0.6282722513089005,
      "grad_norm": 3.7619833946228027,
      "learning_rate": 7.485000000000001e-06,
      "loss": 0.5347,
      "step": 1500
    },
    {
      "epoch": 0.6282722513089005,
      "eval_loss": 0.4724981188774109,
      "eval_runtime": 271.4755,
      "eval_samples_per_second": 31.27,
      "eval_steps_per_second": 3.912,
      "step": 1500
    },
    {
      "epoch": 0.6492146596858639,
      "grad_norm": 2.4959261417388916,
      "learning_rate": 7.735e-06,
      "loss": 0.5247,
      "step": 1550
    },
    {
      "epoch": 0.6701570680628273,
      "grad_norm": 4.8328447341918945,
      "learning_rate": 7.985e-06,
      "loss": 0.5263,
      "step": 1600
    },
    {
      "epoch": 0.6910994764397905,
      "grad_norm": 3.3857789039611816,
      "learning_rate": 8.235e-06,
      "loss": 0.5366,
      "step": 1650
    },
    {
      "epoch": 0.7120418848167539,
      "grad_norm": 6.467920780181885,
      "learning_rate": 8.485000000000001e-06,
      "loss": 0.5233,
      "step": 1700
    },
    {
      "epoch": 0.7329842931937173,
      "grad_norm": 3.377429485321045,
      "learning_rate": 8.735000000000002e-06,
      "loss": 0.528,
      "step": 1750
    },
    {
      "epoch": 0.7539267015706806,
      "grad_norm": 2.8954410552978516,
      "learning_rate": 8.985000000000001e-06,
      "loss": 0.5072,
      "step": 1800
    },
    {
      "epoch": 0.774869109947644,
      "grad_norm": 5.599782943725586,
      "learning_rate": 9.235e-06,
      "loss": 0.5337,
      "step": 1850
    },
    {
      "epoch": 0.7958115183246073,
      "grad_norm": 3.421278238296509,
      "learning_rate": 9.485000000000002e-06,
      "loss": 0.506,
      "step": 1900
    },
    {
      "epoch": 0.8167539267015707,
      "grad_norm": 2.43534517288208,
      "learning_rate": 9.735e-06,
      "loss": 0.503,
      "step": 1950
    },
    {
      "epoch": 0.837696335078534,
      "grad_norm": 3.1177427768707275,
      "learning_rate": 9.985000000000002e-06,
      "loss": 0.4888,
      "step": 2000
    },
    {
      "epoch": 0.837696335078534,
      "eval_loss": 0.46118447184562683,
      "eval_runtime": 279.0953,
      "eval_samples_per_second": 30.416,
      "eval_steps_per_second": 3.805,
      "step": 2000
    },
    {
      "epoch": 0.8586387434554974,
      "grad_norm": 3.640716075897217,
      "learning_rate": 9.765e-06,
      "loss": 0.5128,
      "step": 2050
    },
    {
      "epoch": 0.8795811518324608,
      "grad_norm": 6.894229412078857,
      "learning_rate": 9.515000000000001e-06,
      "loss": 0.4882,
      "step": 2100
    },
    {
      "epoch": 0.900523560209424,
      "grad_norm": 3.4520230293273926,
      "learning_rate": 9.265e-06,
      "loss": 0.4932,
      "step": 2150
    },
    {
      "epoch": 0.9214659685863874,
      "grad_norm": 3.128208875656128,
      "learning_rate": 9.015000000000001e-06,
      "loss": 0.4861,
      "step": 2200
    },
    {
      "epoch": 0.9424083769633508,
      "grad_norm": 3.2562131881713867,
      "learning_rate": 8.765e-06,
      "loss": 0.495,
      "step": 2250
    },
    {
      "epoch": 0.9633507853403142,
      "grad_norm": 4.8225603103637695,
      "learning_rate": 8.515e-06,
      "loss": 0.4968,
      "step": 2300
    },
    {
      "epoch": 0.9842931937172775,
      "grad_norm": 3.6048710346221924,
      "learning_rate": 8.265000000000001e-06,
      "loss": 0.4975,
      "step": 2350
    },
    {
      "epoch": 1.0052356020942408,
      "grad_norm": 5.707277774810791,
      "learning_rate": 8.015e-06,
      "loss": 0.4717,
      "step": 2400
    },
    {
      "epoch": 1.0261780104712042,
      "grad_norm": 3.100660800933838,
      "learning_rate": 7.765000000000001e-06,
      "loss": 0.489,
      "step": 2450
    },
    {
      "epoch": 1.0471204188481675,
      "grad_norm": 2.2402753829956055,
      "learning_rate": 7.515e-06,
      "loss": 0.4923,
      "step": 2500
    },
    {
      "epoch": 1.0471204188481675,
      "eval_loss": 0.4335620403289795,
      "eval_runtime": 275.5568,
      "eval_samples_per_second": 30.807,
      "eval_steps_per_second": 3.854,
      "step": 2500
    },
    {
      "epoch": 1.068062827225131,
      "grad_norm": 4.66270112991333,
      "learning_rate": 7.265000000000001e-06,
      "loss": 0.4845,
      "step": 2550
    },
    {
      "epoch": 1.0890052356020943,
      "grad_norm": 2.6913790702819824,
      "learning_rate": 7.015000000000001e-06,
      "loss": 0.4614,
      "step": 2600
    },
    {
      "epoch": 1.1099476439790577,
      "grad_norm": 2.7639589309692383,
      "learning_rate": 6.7650000000000005e-06,
      "loss": 0.4844,
      "step": 2650
    },
    {
      "epoch": 1.130890052356021,
      "grad_norm": 3.4958367347717285,
      "learning_rate": 6.515e-06,
      "loss": 0.485,
      "step": 2700
    },
    {
      "epoch": 1.1518324607329844,
      "grad_norm": 1.9745889902114868,
      "learning_rate": 6.265e-06,
      "loss": 0.4875,
      "step": 2750
    },
    {
      "epoch": 1.1727748691099475,
      "grad_norm": 2.011784553527832,
      "learning_rate": 6.015000000000001e-06,
      "loss": 0.4637,
      "step": 2800
    },
    {
      "epoch": 1.193717277486911,
      "grad_norm": 4.093133926391602,
      "learning_rate": 5.765000000000001e-06,
      "loss": 0.4665,
      "step": 2850
    },
    {
      "epoch": 1.2146596858638743,
      "grad_norm": 3.1722769737243652,
      "learning_rate": 5.5150000000000006e-06,
      "loss": 0.4708,
      "step": 2900
    },
    {
      "epoch": 1.2356020942408377,
      "grad_norm": 3.777724504470825,
      "learning_rate": 5.265e-06,
      "loss": 0.4697,
      "step": 2950
    },
    {
      "epoch": 1.256544502617801,
      "grad_norm": 2.69818377494812,
      "learning_rate": 5.015e-06,
      "loss": 0.4686,
      "step": 3000
    },
    {
      "epoch": 1.256544502617801,
      "eval_loss": 0.4159424901008606,
      "eval_runtime": 272.6799,
      "eval_samples_per_second": 31.132,
      "eval_steps_per_second": 3.895,
      "step": 3000
    },
    {
      "epoch": 1.2774869109947644,
      "grad_norm": 16.021724700927734,
      "learning_rate": 4.765e-06,
      "loss": 0.4693,
      "step": 3050
    },
    {
      "epoch": 1.2984293193717278,
      "grad_norm": 1.931344985961914,
      "learning_rate": 4.515000000000001e-06,
      "loss": 0.4643,
      "step": 3100
    },
    {
      "epoch": 1.3193717277486912,
      "grad_norm": 2.3644895553588867,
      "learning_rate": 4.265000000000001e-06,
      "loss": 0.4815,
      "step": 3150
    },
    {
      "epoch": 1.3403141361256545,
      "grad_norm": 2.145608901977539,
      "learning_rate": 4.0150000000000005e-06,
      "loss": 0.4664,
      "step": 3200
    },
    {
      "epoch": 1.3612565445026177,
      "grad_norm": 5.124729156494141,
      "learning_rate": 3.7650000000000004e-06,
      "loss": 0.4795,
      "step": 3250
    },
    {
      "epoch": 1.3821989528795813,
      "grad_norm": 1.9456732273101807,
      "learning_rate": 3.5150000000000002e-06,
      "loss": 0.4619,
      "step": 3300
    },
    {
      "epoch": 1.4031413612565444,
      "grad_norm": 1.9677468538284302,
      "learning_rate": 3.2650000000000005e-06,
      "loss": 0.4679,
      "step": 3350
    },
    {
      "epoch": 1.4240837696335078,
      "grad_norm": 3.046672821044922,
      "learning_rate": 3.0150000000000004e-06,
      "loss": 0.4601,
      "step": 3400
    },
    {
      "epoch": 1.4450261780104712,
      "grad_norm": 3.231550455093384,
      "learning_rate": 2.7650000000000006e-06,
      "loss": 0.4648,
      "step": 3450
    },
    {
      "epoch": 1.4659685863874345,
      "grad_norm": 1.9561330080032349,
      "learning_rate": 2.515e-06,
      "loss": 0.4608,
      "step": 3500
    },
    {
      "epoch": 1.4659685863874345,
      "eval_loss": 0.41274404525756836,
      "eval_runtime": 275.8694,
      "eval_samples_per_second": 30.772,
      "eval_steps_per_second": 3.85,
      "step": 3500
    },
    {
      "epoch": 1.486910994764398,
      "grad_norm": 3.744529962539673,
      "learning_rate": 2.2650000000000003e-06,
      "loss": 0.4738,
      "step": 3550
    },
    {
      "epoch": 1.5078534031413613,
      "grad_norm": 2.4467790126800537,
      "learning_rate": 2.015e-06,
      "loss": 0.4626,
      "step": 3600
    },
    {
      "epoch": 1.5287958115183247,
      "grad_norm": 1.932844638824463,
      "learning_rate": 1.765e-06,
      "loss": 0.4706,
      "step": 3650
    },
    {
      "epoch": 1.5497382198952878,
      "grad_norm": 1.966614842414856,
      "learning_rate": 1.5150000000000001e-06,
      "loss": 0.4627,
      "step": 3700
    },
    {
      "epoch": 1.5706806282722514,
      "grad_norm": 2.5016965866088867,
      "learning_rate": 1.2650000000000002e-06,
      "loss": 0.4584,
      "step": 3750
    },
    {
      "epoch": 1.5916230366492146,
      "grad_norm": 2.366603136062622,
      "learning_rate": 1.02e-06,
      "loss": 0.4683,
      "step": 3800
    },
    {
      "epoch": 1.6125654450261782,
      "grad_norm": 1.9269238710403442,
      "learning_rate": 7.7e-07,
      "loss": 0.4863,
      "step": 3850
    },
    {
      "epoch": 1.6335078534031413,
      "grad_norm": 2.289233684539795,
      "learning_rate": 5.2e-07,
      "loss": 0.4594,
      "step": 3900
    },
    {
      "epoch": 1.6544502617801047,
      "grad_norm": 6.539655685424805,
      "learning_rate": 2.7e-07,
      "loss": 0.4643,
      "step": 3950
    },
    {
      "epoch": 1.675392670157068,
      "grad_norm": 2.761108875274658,
      "learning_rate": 2e-08,
      "loss": 0.4567,
      "step": 4000
    },
    {
      "epoch": 1.675392670157068,
      "eval_loss": 0.40972957015037537,
      "eval_runtime": 274.5113,
      "eval_samples_per_second": 30.924,
      "eval_steps_per_second": 3.869,
      "step": 4000
    }
  ],
  "logging_steps": 50,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "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": 1.791435708710208e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}