{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 13196,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06062443164595332,
      "grad_norm": 11.75,
      "learning_rate": 9.998590021800365e-06,
      "loss": 1.3697,
      "memory/device_reserved (GiB)": 4.98,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 200,
      "tokens_per_second_per_gpu": 2604.8
    },
    {
      "epoch": 0.12124886329190664,
      "grad_norm": 11.625,
      "learning_rate": 9.987143623124135e-06,
      "loss": 1.3194,
      "memory/device_reserved (GiB)": 4.98,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 400,
      "tokens_per_second_per_gpu": 2728.55
    },
    {
      "epoch": 0.18187329493785995,
      "grad_norm": 5.1875,
      "learning_rate": 9.964219627493663e-06,
      "loss": 1.3388,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 6.73,
      "memory/max_allocated (GiB)": 6.73,
      "step": 600,
      "tokens_per_second_per_gpu": 1955.81
    },
    {
      "epoch": 0.24249772658381327,
      "grad_norm": 9.1875,
      "learning_rate": 9.929870793041122e-06,
      "loss": 1.3009,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 800,
      "tokens_per_second_per_gpu": 2947.03
    },
    {
      "epoch": 0.30312215822976657,
      "grad_norm": 6.59375,
      "learning_rate": 9.884176171445684e-06,
      "loss": 1.2999,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 1000,
      "tokens_per_second_per_gpu": 1335.52
    },
    {
      "epoch": 0.3637465898757199,
      "grad_norm": 8.8125,
      "learning_rate": 9.82724092600107e-06,
      "loss": 1.2861,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 1200,
      "tokens_per_second_per_gpu": 2361.09
    },
    {
      "epoch": 0.4243710215216732,
      "grad_norm": 9.1875,
      "learning_rate": 9.759196089588842e-06,
      "loss": 1.2967,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 1400,
      "tokens_per_second_per_gpu": 2718.04
    },
    {
      "epoch": 0.48499545316762654,
      "grad_norm": 9.125,
      "learning_rate": 9.68019826311448e-06,
      "loss": 1.2745,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 1600,
      "tokens_per_second_per_gpu": 3015.46
    },
    {
      "epoch": 0.5456198848135799,
      "grad_norm": 8.625,
      "learning_rate": 9.59042925510027e-06,
      "loss": 1.283,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 1800,
      "tokens_per_second_per_gpu": 3223.33
    },
    {
      "epoch": 0.6062443164595331,
      "grad_norm": 9.375,
      "learning_rate": 9.49009566326443e-06,
      "loss": 1.3139,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 2000,
      "tokens_per_second_per_gpu": 2673.11
    },
    {
      "epoch": 0.6668687481054865,
      "grad_norm": 7.71875,
      "learning_rate": 9.37942839904948e-06,
      "loss": 1.3725,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 2200,
      "tokens_per_second_per_gpu": 1867.41
    },
    {
      "epoch": 0.7274931797514398,
      "grad_norm": 7.84375,
      "learning_rate": 9.258682156194094e-06,
      "loss": 1.3003,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 2400,
      "tokens_per_second_per_gpu": 2598.4
    },
    {
      "epoch": 0.7881176113973931,
      "grad_norm": 10.6875,
      "learning_rate": 9.128134824571508e-06,
      "loss": 1.325,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 2600,
      "tokens_per_second_per_gpu": 2643.37
    },
    {
      "epoch": 0.8487420430433464,
      "grad_norm": 11.4375,
      "learning_rate": 8.988086850643474e-06,
      "loss": 1.3052,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 2800,
      "tokens_per_second_per_gpu": 3103.11
    },
    {
      "epoch": 0.9093664746892998,
      "grad_norm": 9.3125,
      "learning_rate": 8.838860546001651e-06,
      "loss": 1.3247,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 3000,
      "tokens_per_second_per_gpu": 2654.71
    },
    {
      "epoch": 0.9699909063352531,
      "grad_norm": 10.875,
      "learning_rate": 8.680799345587778e-06,
      "loss": 1.2945,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 3200,
      "tokens_per_second_per_gpu": 2909.79
    },
    {
      "epoch": 1.0306153379812064,
      "grad_norm": 8.25,
      "learning_rate": 8.514267017299784e-06,
      "loss": 1.2225,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 3400,
      "tokens_per_second_per_gpu": 3081.31
    },
    {
      "epoch": 1.0912397696271596,
      "grad_norm": 7.6875,
      "learning_rate": 8.339646824802882e-06,
      "loss": 1.1383,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 3600,
      "tokens_per_second_per_gpu": 1820.96
    },
    {
      "epoch": 1.151864201273113,
      "grad_norm": 9.125,
      "learning_rate": 8.15734064547238e-06,
      "loss": 1.0931,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 3800,
      "tokens_per_second_per_gpu": 2334.09
    },
    {
      "epoch": 1.2124886329190665,
      "grad_norm": 8.8125,
      "learning_rate": 7.967768045498217e-06,
      "loss": 1.0925,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 4000,
      "tokens_per_second_per_gpu": 2451.13
    },
    {
      "epoch": 1.2731130645650197,
      "grad_norm": 6.90625,
      "learning_rate": 7.771365314279794e-06,
      "loss": 1.0901,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 4200,
      "tokens_per_second_per_gpu": 1827.05
    },
    {
      "epoch": 1.333737496210973,
      "grad_norm": 21.25,
      "learning_rate": 7.568584460333408e-06,
      "loss": 1.1244,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 4400,
      "tokens_per_second_per_gpu": 3097.57
    },
    {
      "epoch": 1.3943619278569264,
      "grad_norm": 8.9375,
      "learning_rate": 7.3598921710231106e-06,
      "loss": 1.1705,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 4600,
      "tokens_per_second_per_gpu": 2846.72
    },
    {
      "epoch": 1.4549863595028796,
      "grad_norm": 18.75,
      "learning_rate": 7.1457687385091415e-06,
      "loss": 1.1692,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 4800,
      "tokens_per_second_per_gpu": 1278.69
    },
    {
      "epoch": 1.515610791148833,
      "grad_norm": 11.875,
      "learning_rate": 6.926706954385761e-06,
      "loss": 1.1225,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 5000,
      "tokens_per_second_per_gpu": 2819.93
    },
    {
      "epoch": 1.5762352227947862,
      "grad_norm": 9.1875,
      "learning_rate": 6.7032109755524384e-06,
      "loss": 1.155,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 5200,
      "tokens_per_second_per_gpu": 2809.42
    },
    {
      "epoch": 1.6368596544407397,
      "grad_norm": 7.96875,
      "learning_rate": 6.475795163928501e-06,
      "loss": 1.153,
      "memory/device_reserved (GiB)": 6.92,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 5400,
      "tokens_per_second_per_gpu": 2595.08
    },
    {
      "epoch": 1.697484086086693,
      "grad_norm": 7.21875,
      "learning_rate": 6.244982902681568e-06,
      "loss": 1.1702,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 6.73,
      "memory/max_allocated (GiB)": 6.73,
      "step": 5600,
      "tokens_per_second_per_gpu": 2620.26
    },
    {
      "epoch": 1.758108517732646,
      "grad_norm": 7.84375,
      "learning_rate": 6.011305391694152e-06,
      "loss": 1.1159,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 5800,
      "tokens_per_second_per_gpu": 2988.73
    },
    {
      "epoch": 1.8187329493785995,
      "grad_norm": 4.96875,
      "learning_rate": 5.775300425040592e-06,
      "loss": 1.1151,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 6000,
      "tokens_per_second_per_gpu": 2139.4
    },
    {
      "epoch": 1.879357381024553,
      "grad_norm": 7.90625,
      "learning_rate": 5.537511153287856e-06,
      "loss": 1.1135,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 6200,
      "tokens_per_second_per_gpu": 3019.1
    },
    {
      "epoch": 1.9399818126705062,
      "grad_norm": 7.96875,
      "learning_rate": 5.298484833468716e-06,
      "loss": 1.1488,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 6400,
      "tokens_per_second_per_gpu": 2381.51
    },
    {
      "epoch": 2.0006062443164594,
      "grad_norm": 10.4375,
      "learning_rate": 5.0587715696041685e-06,
      "loss": 1.1651,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 6600,
      "tokens_per_second_per_gpu": 3025.08
    },
    {
      "epoch": 2.061230675962413,
      "grad_norm": 9.0,
      "learning_rate": 4.8189230466736585e-06,
      "loss": 1.0324,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 6800,
      "tokens_per_second_per_gpu": 2687.95
    },
    {
      "epoch": 2.1218551076083663,
      "grad_norm": 8.9375,
      "learning_rate": 4.579491260946856e-06,
      "loss": 0.9765,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 7000,
      "tokens_per_second_per_gpu": 2745.94
    },
    {
      "epoch": 2.1824795392543193,
      "grad_norm": 9.875,
      "learning_rate": 4.341027249598999e-06,
      "loss": 1.0136,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 7200,
      "tokens_per_second_per_gpu": 2545.45
    },
    {
      "epoch": 2.2431039709002727,
      "grad_norm": 8.1875,
      "learning_rate": 4.104079822533548e-06,
      "loss": 1.0314,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 6.73,
      "memory/max_allocated (GiB)": 6.73,
      "step": 7400,
      "tokens_per_second_per_gpu": 3023.1
    },
    {
      "epoch": 2.303728402546226,
      "grad_norm": 7.90625,
      "learning_rate": 3.869194299330752e-06,
      "loss": 1.0273,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 7600,
      "tokens_per_second_per_gpu": 2507.53
    },
    {
      "epoch": 2.3643528341921796,
      "grad_norm": 8.3125,
      "learning_rate": 3.63691125422898e-06,
      "loss": 1.0177,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 7800,
      "tokens_per_second_per_gpu": 2346.84
    },
    {
      "epoch": 2.424977265838133,
      "grad_norm": 43.75,
      "learning_rate": 3.407765272027156e-06,
      "loss": 1.0256,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 8000,
      "tokens_per_second_per_gpu": 2946.41
    },
    {
      "epoch": 2.485601697484086,
      "grad_norm": 8.9375,
      "learning_rate": 3.1822837177715192e-06,
      "loss": 1.0005,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 8200,
      "tokens_per_second_per_gpu": 2745.12
    },
    {
      "epoch": 2.5462261291300394,
      "grad_norm": 10.375,
      "learning_rate": 2.960985523058174e-06,
      "loss": 1.0119,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 8400,
      "tokens_per_second_per_gpu": 2990.09
    },
    {
      "epoch": 2.606850560775993,
      "grad_norm": 5.3125,
      "learning_rate": 2.744379991744713e-06,
      "loss": 1.0293,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 8600,
      "tokens_per_second_per_gpu": 2229.96
    },
    {
      "epoch": 2.667474992421946,
      "grad_norm": 6.75,
      "learning_rate": 2.5329656278194625e-06,
      "loss": 1.0065,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 8800,
      "tokens_per_second_per_gpu": 1817.5
    },
    {
      "epoch": 2.7280994240678993,
      "grad_norm": 9.75,
      "learning_rate": 2.327228988125946e-06,
      "loss": 1.0057,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 9000,
      "tokens_per_second_per_gpu": 2842.21
    },
    {
      "epoch": 2.7887238557138527,
      "grad_norm": 9.5625,
      "learning_rate": 2.1276435625829543e-06,
      "loss": 0.9829,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 9200,
      "tokens_per_second_per_gpu": 3086.49
    },
    {
      "epoch": 2.849348287359806,
      "grad_norm": 9.5,
      "learning_rate": 1.9346686844773253e-06,
      "loss": 1.0178,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 9400,
      "tokens_per_second_per_gpu": 2586.56
    },
    {
      "epoch": 2.909972719005759,
      "grad_norm": 7.53125,
      "learning_rate": 1.7487484733373017e-06,
      "loss": 1.0099,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 9600,
      "tokens_per_second_per_gpu": 1847.53
    },
    {
      "epoch": 2.9705971506517126,
      "grad_norm": 12.5,
      "learning_rate": 1.5703108128194145e-06,
      "loss": 1.0223,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 9800,
      "tokens_per_second_per_gpu": 2763.65
    },
    {
      "epoch": 3.031221582297666,
      "grad_norm": 8.0625,
      "learning_rate": 1.3997663659611982e-06,
      "loss": 0.9828,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 10000,
      "tokens_per_second_per_gpu": 2827.68
    },
    {
      "epoch": 3.0918460139436195,
      "grad_norm": 11.75,
      "learning_rate": 1.2375076300660677e-06,
      "loss": 0.9693,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 10200,
      "tokens_per_second_per_gpu": 2538.93
    },
    {
      "epoch": 3.1524704455895725,
      "grad_norm": 4.78125,
      "learning_rate": 1.083908033395496e-06,
      "loss": 0.9715,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 10400,
      "tokens_per_second_per_gpu": 2718.63
    },
    {
      "epoch": 3.213094877235526,
      "grad_norm": 9.75,
      "learning_rate": 9.393210757473959e-07,
      "loss": 1.0057,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 10600,
      "tokens_per_second_per_gpu": 2702.32
    },
    {
      "epoch": 3.2737193088814793,
      "grad_norm": 8.25,
      "learning_rate": 8.040795148985875e-07,
      "loss": 0.9761,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 10800,
      "tokens_per_second_per_gpu": 2937.95
    },
    {
      "epoch": 3.3343437405274328,
      "grad_norm": 7.53125,
      "learning_rate": 6.784946007837395e-07,
      "loss": 0.9284,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 11000,
      "tokens_per_second_per_gpu": 2385.64
    },
    {
      "epoch": 3.3949681721733858,
      "grad_norm": 6.875,
      "learning_rate": 5.628553591732478e-07,
      "loss": 0.9631,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 5.61,
      "memory/max_allocated (GiB)": 5.61,
      "step": 11200,
      "tokens_per_second_per_gpu": 2425.36
    },
    {
      "epoch": 3.455592603819339,
      "grad_norm": 8.0625,
      "learning_rate": 4.57427926498627e-07,
      "loss": 0.986,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 11400,
      "tokens_per_second_per_gpu": 2226.29
    },
    {
      "epoch": 3.5162170354652926,
      "grad_norm": 9.875,
      "learning_rate": 3.6245493735626913e-07,
      "loss": 1.0002,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 11600,
      "tokens_per_second_per_gpu": 2925.67
    },
    {
      "epoch": 3.5768414671112456,
      "grad_norm": 9.8125,
      "learning_rate": 2.7815496609921053e-07,
      "loss": 0.9356,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 11800,
      "tokens_per_second_per_gpu": 2927.99
    },
    {
      "epoch": 3.637465898757199,
      "grad_norm": 11.4375,
      "learning_rate": 2.047220238020181e-07,
      "loss": 1.0038,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 12000,
      "tokens_per_second_per_gpu": 2644.97
    },
    {
      "epoch": 3.6980903304031525,
      "grad_norm": 8.6875,
      "learning_rate": 1.4232511175652108e-07,
      "loss": 0.9089,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 12200,
      "tokens_per_second_per_gpu": 1991.61
    },
    {
      "epoch": 3.758714762049106,
      "grad_norm": 9.8125,
      "learning_rate": 9.110783252598453e-08,
      "loss": 0.9737,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 12400,
      "tokens_per_second_per_gpu": 2470.36
    },
    {
      "epoch": 3.8193391936950594,
      "grad_norm": 38.0,
      "learning_rate": 5.118805945285188e-08,
      "loss": 0.9549,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 12600,
      "tokens_per_second_per_gpu": 3157.65
    },
    {
      "epoch": 3.8799636253410124,
      "grad_norm": 10.125,
      "learning_rate": 2.2657665380669824e-08,
      "loss": 0.9517,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 12800,
      "tokens_per_second_per_gpu": 2685.23
    },
    {
      "epoch": 3.940588056986966,
      "grad_norm": 11.875,
      "learning_rate": 5.5823112145286796e-09,
      "loss": 0.9891,
      "memory/device_reserved (GiB)": 6.96,
      "memory/max_active (GiB)": 4.63,
      "memory/max_allocated (GiB)": 4.63,
      "step": 13000,
      "tokens_per_second_per_gpu": 2045.68
    }
  ],
  "logging_steps": 200,
  "max_steps": 13196,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 3299,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 5.304147161540198e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}