{
  "best_metric": 0.774284303188324,
  "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l58-l/checkpoint-21500",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 21712,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09209799226376865,
      "grad_norm": 0.0014697719598188996,
      "learning_rate": 4.884877509670289e-08,
      "loss": 0.4588,
      "step": 500
    },
    {
      "epoch": 0.09209799226376865,
      "eval_loss": 1.4194883108139038,
      "eval_runtime": 78.7703,
      "eval_samples_per_second": 15.31,
      "eval_steps_per_second": 1.917,
      "step": 500
    },
    {
      "epoch": 0.1841959845275373,
      "grad_norm": 301.76043701171875,
      "learning_rate": 4.769755019340578e-08,
      "loss": 0.4255,
      "step": 1000
    },
    {
      "epoch": 0.1841959845275373,
      "eval_loss": 1.3416553735733032,
      "eval_runtime": 77.0051,
      "eval_samples_per_second": 15.661,
      "eval_steps_per_second": 1.961,
      "step": 1000
    },
    {
      "epoch": 0.27629397679130596,
      "grad_norm": 35.28853225708008,
      "learning_rate": 4.654632529010867e-08,
      "loss": 0.3724,
      "step": 1500
    },
    {
      "epoch": 0.27629397679130596,
      "eval_loss": 1.2872973680496216,
      "eval_runtime": 77.5439,
      "eval_samples_per_second": 15.552,
      "eval_steps_per_second": 1.947,
      "step": 1500
    },
    {
      "epoch": 0.3683919690550746,
      "grad_norm": 338.9349060058594,
      "learning_rate": 4.539510038681156e-08,
      "loss": 0.3251,
      "step": 2000
    },
    {
      "epoch": 0.3683919690550746,
      "eval_loss": 1.234910249710083,
      "eval_runtime": 78.7715,
      "eval_samples_per_second": 15.31,
      "eval_steps_per_second": 1.917,
      "step": 2000
    },
    {
      "epoch": 0.46048996131884323,
      "grad_norm": 325.85101318359375,
      "learning_rate": 4.4243875483514457e-08,
      "loss": 0.3308,
      "step": 2500
    },
    {
      "epoch": 0.46048996131884323,
      "eval_loss": 1.1944907903671265,
      "eval_runtime": 79.0397,
      "eval_samples_per_second": 15.258,
      "eval_steps_per_second": 1.91,
      "step": 2500
    },
    {
      "epoch": 0.5525879535826119,
      "grad_norm": 100.85855102539062,
      "learning_rate": 4.309265058021735e-08,
      "loss": 0.3017,
      "step": 3000
    },
    {
      "epoch": 0.5525879535826119,
      "eval_loss": 1.1593009233474731,
      "eval_runtime": 77.6354,
      "eval_samples_per_second": 15.534,
      "eval_steps_per_second": 1.945,
      "step": 3000
    },
    {
      "epoch": 0.6446859458463805,
      "grad_norm": 8.865598678588867,
      "learning_rate": 4.194142567692024e-08,
      "loss": 0.2962,
      "step": 3500
    },
    {
      "epoch": 0.6446859458463805,
      "eval_loss": 1.1259427070617676,
      "eval_runtime": 78.2715,
      "eval_samples_per_second": 15.408,
      "eval_steps_per_second": 1.929,
      "step": 3500
    },
    {
      "epoch": 0.7367839381101492,
      "grad_norm": 4.393447399139404,
      "learning_rate": 4.079020077362314e-08,
      "loss": 0.2919,
      "step": 4000
    },
    {
      "epoch": 0.7367839381101492,
      "eval_loss": 1.0954023599624634,
      "eval_runtime": 78.5366,
      "eval_samples_per_second": 15.356,
      "eval_steps_per_second": 1.923,
      "step": 4000
    },
    {
      "epoch": 0.8288819303739179,
      "grad_norm": 1.6036031246185303,
      "learning_rate": 3.9638975870326026e-08,
      "loss": 0.307,
      "step": 4500
    },
    {
      "epoch": 0.8288819303739179,
      "eval_loss": 1.0728861093521118,
      "eval_runtime": 78.4032,
      "eval_samples_per_second": 15.382,
      "eval_steps_per_second": 1.926,
      "step": 4500
    },
    {
      "epoch": 0.9209799226376865,
      "grad_norm": 11.990758895874023,
      "learning_rate": 3.8487750967028915e-08,
      "loss": 0.2764,
      "step": 5000
    },
    {
      "epoch": 0.9209799226376865,
      "eval_loss": 1.0523593425750732,
      "eval_runtime": 78.3726,
      "eval_samples_per_second": 15.388,
      "eval_steps_per_second": 1.927,
      "step": 5000
    },
    {
      "epoch": 1.013077914901455,
      "grad_norm": 43.865562438964844,
      "learning_rate": 3.733652606373181e-08,
      "loss": 0.2456,
      "step": 5500
    },
    {
      "epoch": 1.013077914901455,
      "eval_loss": 1.0375442504882812,
      "eval_runtime": 78.2473,
      "eval_samples_per_second": 15.413,
      "eval_steps_per_second": 1.93,
      "step": 5500
    },
    {
      "epoch": 1.1051759071652238,
      "grad_norm": 105.91036987304688,
      "learning_rate": 3.61853011604347e-08,
      "loss": 0.2642,
      "step": 6000
    },
    {
      "epoch": 1.1051759071652238,
      "eval_loss": 1.0233356952667236,
      "eval_runtime": 78.3135,
      "eval_samples_per_second": 15.4,
      "eval_steps_per_second": 1.928,
      "step": 6000
    },
    {
      "epoch": 1.1972738994289924,
      "grad_norm": 221.4896240234375,
      "learning_rate": 3.503407625713759e-08,
      "loss": 0.2066,
      "step": 6500
    },
    {
      "epoch": 1.1972738994289924,
      "eval_loss": 1.0104238986968994,
      "eval_runtime": 78.6485,
      "eval_samples_per_second": 15.334,
      "eval_steps_per_second": 1.92,
      "step": 6500
    },
    {
      "epoch": 1.289371891692761,
      "grad_norm": 270.53253173828125,
      "learning_rate": 3.3882851353840485e-08,
      "loss": 0.2376,
      "step": 7000
    },
    {
      "epoch": 1.289371891692761,
      "eval_loss": 0.998353898525238,
      "eval_runtime": 78.5502,
      "eval_samples_per_second": 15.353,
      "eval_steps_per_second": 1.922,
      "step": 7000
    },
    {
      "epoch": 1.3814698839565298,
      "grad_norm": 402.9339904785156,
      "learning_rate": 3.2731626450543374e-08,
      "loss": 0.1931,
      "step": 7500
    },
    {
      "epoch": 1.3814698839565298,
      "eval_loss": 0.9887453317642212,
      "eval_runtime": 78.4833,
      "eval_samples_per_second": 15.366,
      "eval_steps_per_second": 1.924,
      "step": 7500
    },
    {
      "epoch": 1.4735678762202984,
      "grad_norm": 0.006355441175401211,
      "learning_rate": 3.158040154724626e-08,
      "loss": 0.2163,
      "step": 8000
    },
    {
      "epoch": 1.4735678762202984,
      "eval_loss": 0.9767189621925354,
      "eval_runtime": 78.5499,
      "eval_samples_per_second": 15.353,
      "eval_steps_per_second": 1.922,
      "step": 8000
    },
    {
      "epoch": 1.565665868484067,
      "grad_norm": 0.12982100248336792,
      "learning_rate": 3.042917664394916e-08,
      "loss": 0.1903,
      "step": 8500
    },
    {
      "epoch": 1.565665868484067,
      "eval_loss": 0.9664921760559082,
      "eval_runtime": 78.6832,
      "eval_samples_per_second": 15.327,
      "eval_steps_per_second": 1.919,
      "step": 8500
    },
    {
      "epoch": 1.6577638607478358,
      "grad_norm": 173.31788635253906,
      "learning_rate": 2.927795174065205e-08,
      "loss": 0.2069,
      "step": 9000
    },
    {
      "epoch": 1.6577638607478358,
      "eval_loss": 0.9571623206138611,
      "eval_runtime": 78.544,
      "eval_samples_per_second": 15.354,
      "eval_steps_per_second": 1.922,
      "step": 9000
    },
    {
      "epoch": 1.7498618530116044,
      "grad_norm": 358.7403259277344,
      "learning_rate": 2.8126726837354947e-08,
      "loss": 0.2093,
      "step": 9500
    },
    {
      "epoch": 1.7498618530116044,
      "eval_loss": 0.9496769905090332,
      "eval_runtime": 78.3124,
      "eval_samples_per_second": 15.4,
      "eval_steps_per_second": 1.928,
      "step": 9500
    },
    {
      "epoch": 1.841959845275373,
      "grad_norm": 435.9449768066406,
      "learning_rate": 2.6975501934057836e-08,
      "loss": 0.2523,
      "step": 10000
    },
    {
      "epoch": 1.841959845275373,
      "eval_loss": 0.9420022368431091,
      "eval_runtime": 78.3702,
      "eval_samples_per_second": 15.388,
      "eval_steps_per_second": 1.927,
      "step": 10000
    },
    {
      "epoch": 1.9340578375391417,
      "grad_norm": 23.49118423461914,
      "learning_rate": 2.582427703076073e-08,
      "loss": 0.2127,
      "step": 10500
    },
    {
      "epoch": 1.9340578375391417,
      "eval_loss": 0.9328628182411194,
      "eval_runtime": 76.6904,
      "eval_samples_per_second": 15.726,
      "eval_steps_per_second": 1.969,
      "step": 10500
    },
    {
      "epoch": 2.02615582980291,
      "grad_norm": 0.4029064476490021,
      "learning_rate": 2.467305212746362e-08,
      "loss": 0.1968,
      "step": 11000
    },
    {
      "epoch": 2.02615582980291,
      "eval_loss": 0.926975429058075,
      "eval_runtime": 76.7161,
      "eval_samples_per_second": 15.72,
      "eval_steps_per_second": 1.968,
      "step": 11000
    },
    {
      "epoch": 2.118253822066679,
      "grad_norm": 274.0943603515625,
      "learning_rate": 2.3521827224166513e-08,
      "loss": 0.1879,
      "step": 11500
    },
    {
      "epoch": 2.118253822066679,
      "eval_loss": 0.9231382012367249,
      "eval_runtime": 78.6224,
      "eval_samples_per_second": 15.339,
      "eval_steps_per_second": 1.921,
      "step": 11500
    },
    {
      "epoch": 2.2103518143304477,
      "grad_norm": 0.000765918695833534,
      "learning_rate": 2.2370602320869402e-08,
      "loss": 0.1981,
      "step": 12000
    },
    {
      "epoch": 2.2103518143304477,
      "eval_loss": 0.9183884263038635,
      "eval_runtime": 76.7707,
      "eval_samples_per_second": 15.709,
      "eval_steps_per_second": 1.967,
      "step": 12000
    },
    {
      "epoch": 2.302449806594216,
      "grad_norm": 0.6250036358833313,
      "learning_rate": 2.1219377417572295e-08,
      "loss": 0.1964,
      "step": 12500
    },
    {
      "epoch": 2.302449806594216,
      "eval_loss": 0.9134928584098816,
      "eval_runtime": 78.5128,
      "eval_samples_per_second": 15.361,
      "eval_steps_per_second": 1.923,
      "step": 12500
    },
    {
      "epoch": 2.394547798857985,
      "grad_norm": 158.6990203857422,
      "learning_rate": 2.0068152514275187e-08,
      "loss": 0.1697,
      "step": 13000
    },
    {
      "epoch": 2.394547798857985,
      "eval_loss": 0.910048246383667,
      "eval_runtime": 76.7476,
      "eval_samples_per_second": 15.714,
      "eval_steps_per_second": 1.967,
      "step": 13000
    },
    {
      "epoch": 2.4866457911217537,
      "grad_norm": 1.288072943687439,
      "learning_rate": 1.891692761097808e-08,
      "loss": 0.2015,
      "step": 13500
    },
    {
      "epoch": 2.4866457911217537,
      "eval_loss": 0.9051916599273682,
      "eval_runtime": 78.3909,
      "eval_samples_per_second": 15.384,
      "eval_steps_per_second": 1.926,
      "step": 13500
    },
    {
      "epoch": 2.578743783385522,
      "grad_norm": 5.835799311171286e-05,
      "learning_rate": 1.7765702707680972e-08,
      "loss": 0.1827,
      "step": 14000
    },
    {
      "epoch": 2.578743783385522,
      "eval_loss": 0.9026296734809875,
      "eval_runtime": 76.6591,
      "eval_samples_per_second": 15.732,
      "eval_steps_per_second": 1.97,
      "step": 14000
    },
    {
      "epoch": 2.670841775649291,
      "grad_norm": 30.217784881591797,
      "learning_rate": 1.6614477804383865e-08,
      "loss": 0.1435,
      "step": 14500
    },
    {
      "epoch": 2.670841775649291,
      "eval_loss": 0.899847686290741,
      "eval_runtime": 76.742,
      "eval_samples_per_second": 15.715,
      "eval_steps_per_second": 1.968,
      "step": 14500
    },
    {
      "epoch": 2.7629397679130596,
      "grad_norm": 202.30496215820312,
      "learning_rate": 1.5463252901086757e-08,
      "loss": 0.1541,
      "step": 15000
    },
    {
      "epoch": 2.7629397679130596,
      "eval_loss": 0.896297812461853,
      "eval_runtime": 78.5539,
      "eval_samples_per_second": 15.353,
      "eval_steps_per_second": 1.922,
      "step": 15000
    },
    {
      "epoch": 2.8550377601768284,
      "grad_norm": 241.57069396972656,
      "learning_rate": 1.4312027997789646e-08,
      "loss": 0.1716,
      "step": 15500
    },
    {
      "epoch": 2.8550377601768284,
      "eval_loss": 0.8935310244560242,
      "eval_runtime": 76.9195,
      "eval_samples_per_second": 15.679,
      "eval_steps_per_second": 1.963,
      "step": 15500
    },
    {
      "epoch": 2.947135752440597,
      "grad_norm": 0.25739413499832153,
      "learning_rate": 1.316080309449254e-08,
      "loss": 0.2056,
      "step": 16000
    },
    {
      "epoch": 2.947135752440597,
      "eval_loss": 0.890542209148407,
      "eval_runtime": 76.7268,
      "eval_samples_per_second": 15.718,
      "eval_steps_per_second": 1.968,
      "step": 16000
    },
    {
      "epoch": 3.0392337447043656,
      "grad_norm": 86.5535659790039,
      "learning_rate": 1.2009578191195431e-08,
      "loss": 0.1843,
      "step": 16500
    },
    {
      "epoch": 3.0392337447043656,
      "eval_loss": 0.8875145316123962,
      "eval_runtime": 76.4556,
      "eval_samples_per_second": 15.774,
      "eval_steps_per_second": 1.975,
      "step": 16500
    },
    {
      "epoch": 3.131331736968134,
      "grad_norm": 0.05445564165711403,
      "learning_rate": 1.0858353287898323e-08,
      "loss": 0.1611,
      "step": 17000
    },
    {
      "epoch": 3.131331736968134,
      "eval_loss": 0.8858152627944946,
      "eval_runtime": 77.4814,
      "eval_samples_per_second": 15.565,
      "eval_steps_per_second": 1.949,
      "step": 17000
    },
    {
      "epoch": 3.2240235814296243,
      "grad_norm": 44.673702239990234,
      "learning_rate": 9.699705232129698e-09,
      "loss": 0.1568,
      "step": 17500
    },
    {
      "epoch": 3.2240235814296243,
      "eval_loss": 0.7821062207221985,
      "eval_runtime": 75.6502,
      "eval_samples_per_second": 15.955,
      "eval_steps_per_second": 1.996,
      "step": 17500
    },
    {
      "epoch": 3.316138540899042,
      "grad_norm": 0.5617901086807251,
      "learning_rate": 8.548268238761974e-09,
      "loss": 0.1395,
      "step": 18000
    },
    {
      "epoch": 3.316138540899042,
      "eval_loss": 0.7793659567832947,
      "eval_runtime": 76.0283,
      "eval_samples_per_second": 15.876,
      "eval_steps_per_second": 1.986,
      "step": 18000
    },
    {
      "epoch": 3.4082535003684598,
      "grad_norm": 6.4552321434021,
      "learning_rate": 7.396831245394252e-09,
      "loss": 0.1804,
      "step": 18500
    },
    {
      "epoch": 3.4082535003684598,
      "eval_loss": 0.7778191566467285,
      "eval_runtime": 76.0536,
      "eval_samples_per_second": 15.87,
      "eval_steps_per_second": 1.985,
      "step": 18500
    },
    {
      "epoch": 3.500368459837878,
      "grad_norm": 0.31470850110054016,
      "learning_rate": 6.2453942520265294e-09,
      "loss": 0.1728,
      "step": 19000
    },
    {
      "epoch": 3.500368459837878,
      "eval_loss": 0.7768829464912415,
      "eval_runtime": 76.116,
      "eval_samples_per_second": 15.857,
      "eval_steps_per_second": 1.984,
      "step": 19000
    },
    {
      "epoch": 3.5924834193072956,
      "grad_norm": 0.029128307476639748,
      "learning_rate": 5.093957258658806e-09,
      "loss": 0.179,
      "step": 19500
    },
    {
      "epoch": 3.5924834193072956,
      "eval_loss": 0.7758385539054871,
      "eval_runtime": 76.1889,
      "eval_samples_per_second": 15.842,
      "eval_steps_per_second": 1.982,
      "step": 19500
    },
    {
      "epoch": 3.6845983787767134,
      "grad_norm": 0.0002703067730180919,
      "learning_rate": 3.942520265291083e-09,
      "loss": 0.179,
      "step": 20000
    },
    {
      "epoch": 3.6845983787767134,
      "eval_loss": 0.7751660943031311,
      "eval_runtime": 76.1566,
      "eval_samples_per_second": 15.849,
      "eval_steps_per_second": 1.983,
      "step": 20000
    },
    {
      "epoch": 3.776713338246131,
      "grad_norm": 259.0263366699219,
      "learning_rate": 2.79108327192336e-09,
      "loss": 0.1454,
      "step": 20500
    },
    {
      "epoch": 3.776713338246131,
      "eval_loss": 0.7746543884277344,
      "eval_runtime": 76.3059,
      "eval_samples_per_second": 15.818,
      "eval_steps_per_second": 1.979,
      "step": 20500
    },
    {
      "epoch": 3.868828297715549,
      "grad_norm": 0.0011778937187045813,
      "learning_rate": 1.6396462785556374e-09,
      "loss": 0.1568,
      "step": 21000
    },
    {
      "epoch": 3.868828297715549,
      "eval_loss": 0.7744118571281433,
      "eval_runtime": 76.0825,
      "eval_samples_per_second": 15.864,
      "eval_steps_per_second": 1.985,
      "step": 21000
    },
    {
      "epoch": 3.960943257184967,
      "grad_norm": 0.013251720927655697,
      "learning_rate": 4.882092851879146e-10,
      "loss": 0.1663,
      "step": 21500
    },
    {
      "epoch": 3.960943257184967,
      "eval_loss": 0.774284303188324,
      "eval_runtime": 76.2442,
      "eval_samples_per_second": 15.831,
      "eval_steps_per_second": 1.98,
      "step": 21500
    },
    {
      "epoch": 4.0,
      "step": 21712,
      "total_flos": 7811253201732750.0,
      "train_loss": 0.03522617665827055,
      "train_runtime": 3617.1108,
      "train_samples_per_second": 12.005,
      "train_steps_per_second": 6.003
    }
  ],
  "logging_steps": 500,
  "max_steps": 21712,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "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": 7811253201732750.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}