{
  "best_metric": 0.8749921321868896,
  "best_model_checkpoint": "./output/clip-finetuned-csu-p14-336-e4l58-l/checkpoint-17500",
  "epoch": 3.2240235814296243,
  "eval_steps": 500,
  "global_step": 17500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09211495946941783,
      "grad_norm": 20.28152847290039,
      "learning_rate": 4.8848563006632274e-08,
      "loss": 0.3758,
      "step": 500
    },
    {
      "epoch": 0.09211495946941783,
      "eval_loss": 1.418367862701416,
      "eval_runtime": 73.4439,
      "eval_samples_per_second": 16.434,
      "eval_steps_per_second": 2.056,
      "step": 500
    },
    {
      "epoch": 0.18422991893883567,
      "grad_norm": 47.29538345336914,
      "learning_rate": 4.7697126013264556e-08,
      "loss": 0.4104,
      "step": 1000
    },
    {
      "epoch": 0.18422991893883567,
      "eval_loss": 1.3491536378860474,
      "eval_runtime": 74.3404,
      "eval_samples_per_second": 16.236,
      "eval_steps_per_second": 2.031,
      "step": 1000
    },
    {
      "epoch": 0.2763448784082535,
      "grad_norm": 0.5450155138969421,
      "learning_rate": 4.6545689019896826e-08,
      "loss": 0.4325,
      "step": 1500
    },
    {
      "epoch": 0.2763448784082535,
      "eval_loss": 1.2877168655395508,
      "eval_runtime": 75.3303,
      "eval_samples_per_second": 16.023,
      "eval_steps_per_second": 2.005,
      "step": 1500
    },
    {
      "epoch": 0.36845983787767134,
      "grad_norm": 0.11449883133172989,
      "learning_rate": 4.539425202652911e-08,
      "loss": 0.3423,
      "step": 2000
    },
    {
      "epoch": 0.36845983787767134,
      "eval_loss": 1.2384389638900757,
      "eval_runtime": 76.113,
      "eval_samples_per_second": 15.858,
      "eval_steps_per_second": 1.984,
      "step": 2000
    },
    {
      "epoch": 0.46057479734708917,
      "grad_norm": 389.7380065917969,
      "learning_rate": 4.4242815033161385e-08,
      "loss": 0.3645,
      "step": 2500
    },
    {
      "epoch": 0.46057479734708917,
      "eval_loss": 1.1896915435791016,
      "eval_runtime": 76.23,
      "eval_samples_per_second": 15.834,
      "eval_steps_per_second": 1.981,
      "step": 2500
    },
    {
      "epoch": 0.552689756816507,
      "grad_norm": 340.1103820800781,
      "learning_rate": 4.309137803979366e-08,
      "loss": 0.3171,
      "step": 3000
    },
    {
      "epoch": 0.552689756816507,
      "eval_loss": 1.149291753768921,
      "eval_runtime": 76.1662,
      "eval_samples_per_second": 15.847,
      "eval_steps_per_second": 1.983,
      "step": 3000
    },
    {
      "epoch": 0.6448047162859248,
      "grad_norm": 196.7362518310547,
      "learning_rate": 4.193994104642594e-08,
      "loss": 0.275,
      "step": 3500
    },
    {
      "epoch": 0.6448047162859248,
      "eval_loss": 1.1162768602371216,
      "eval_runtime": 74.6569,
      "eval_samples_per_second": 16.167,
      "eval_steps_per_second": 2.023,
      "step": 3500
    },
    {
      "epoch": 0.7369196757553427,
      "grad_norm": 16.604211807250977,
      "learning_rate": 4.078850405305821e-08,
      "loss": 0.2907,
      "step": 4000
    },
    {
      "epoch": 0.7369196757553427,
      "eval_loss": 1.091786503791809,
      "eval_runtime": 76.3333,
      "eval_samples_per_second": 15.812,
      "eval_steps_per_second": 1.978,
      "step": 4000
    },
    {
      "epoch": 0.8290346352247605,
      "grad_norm": 527.0987548828125,
      "learning_rate": 3.9637067059690496e-08,
      "loss": 0.2581,
      "step": 4500
    },
    {
      "epoch": 0.8290346352247605,
      "eval_loss": 1.0729899406433105,
      "eval_runtime": 74.4912,
      "eval_samples_per_second": 16.203,
      "eval_steps_per_second": 2.027,
      "step": 4500
    },
    {
      "epoch": 0.9211495946941783,
      "grad_norm": 1.148181676864624,
      "learning_rate": 3.848563006632277e-08,
      "loss": 0.268,
      "step": 5000
    },
    {
      "epoch": 0.9211495946941783,
      "eval_loss": 1.052911400794983,
      "eval_runtime": 74.9586,
      "eval_samples_per_second": 16.102,
      "eval_steps_per_second": 2.014,
      "step": 5000
    },
    {
      "epoch": 1.013264554163596,
      "grad_norm": 0.1680648922920227,
      "learning_rate": 3.733419307295505e-08,
      "loss": 0.2467,
      "step": 5500
    },
    {
      "epoch": 1.013264554163596,
      "eval_loss": 1.036412239074707,
      "eval_runtime": 74.707,
      "eval_samples_per_second": 16.156,
      "eval_steps_per_second": 2.021,
      "step": 5500
    },
    {
      "epoch": 1.105379513633014,
      "grad_norm": 269.06585693359375,
      "learning_rate": 3.6182756079587324e-08,
      "loss": 0.2594,
      "step": 6000
    },
    {
      "epoch": 1.105379513633014,
      "eval_loss": 1.0190415382385254,
      "eval_runtime": 74.5657,
      "eval_samples_per_second": 16.187,
      "eval_steps_per_second": 2.025,
      "step": 6000
    },
    {
      "epoch": 1.1974944731024317,
      "grad_norm": 457.15850830078125,
      "learning_rate": 3.50313190862196e-08,
      "loss": 0.2153,
      "step": 6500
    },
    {
      "epoch": 1.1974944731024317,
      "eval_loss": 1.0054280757904053,
      "eval_runtime": 74.5269,
      "eval_samples_per_second": 16.195,
      "eval_steps_per_second": 2.026,
      "step": 6500
    },
    {
      "epoch": 1.2896094325718497,
      "grad_norm": 0.5831326842308044,
      "learning_rate": 3.387988209285188e-08,
      "loss": 0.1792,
      "step": 7000
    },
    {
      "epoch": 1.2896094325718497,
      "eval_loss": 0.9938780665397644,
      "eval_runtime": 74.5283,
      "eval_samples_per_second": 16.195,
      "eval_steps_per_second": 2.026,
      "step": 7000
    },
    {
      "epoch": 1.3817243920412676,
      "grad_norm": 54.12948226928711,
      "learning_rate": 3.272844509948415e-08,
      "loss": 0.208,
      "step": 7500
    },
    {
      "epoch": 1.3817243920412676,
      "eval_loss": 0.9816781282424927,
      "eval_runtime": 76.244,
      "eval_samples_per_second": 15.831,
      "eval_steps_per_second": 1.98,
      "step": 7500
    },
    {
      "epoch": 1.4738393515106853,
      "grad_norm": 0.09778845310211182,
      "learning_rate": 3.1577008106116435e-08,
      "loss": 0.1955,
      "step": 8000
    },
    {
      "epoch": 1.4738393515106853,
      "eval_loss": 0.9722194671630859,
      "eval_runtime": 74.4255,
      "eval_samples_per_second": 16.218,
      "eval_steps_per_second": 2.029,
      "step": 8000
    },
    {
      "epoch": 1.565954310980103,
      "grad_norm": 43.05193328857422,
      "learning_rate": 3.042557111274871e-08,
      "loss": 0.1943,
      "step": 8500
    },
    {
      "epoch": 1.565954310980103,
      "eval_loss": 0.962623119354248,
      "eval_runtime": 74.5615,
      "eval_samples_per_second": 16.188,
      "eval_steps_per_second": 2.025,
      "step": 8500
    },
    {
      "epoch": 1.658069270449521,
      "grad_norm": 0.050531528890132904,
      "learning_rate": 2.927413411938099e-08,
      "loss": 0.2421,
      "step": 9000
    },
    {
      "epoch": 1.658069270449521,
      "eval_loss": 0.9543817043304443,
      "eval_runtime": 76.2957,
      "eval_samples_per_second": 15.82,
      "eval_steps_per_second": 1.979,
      "step": 9000
    },
    {
      "epoch": 1.750184229918939,
      "grad_norm": 0.07409189641475677,
      "learning_rate": 2.8122697126013263e-08,
      "loss": 0.2109,
      "step": 9500
    },
    {
      "epoch": 1.750184229918939,
      "eval_loss": 0.9468172192573547,
      "eval_runtime": 74.551,
      "eval_samples_per_second": 16.19,
      "eval_steps_per_second": 2.025,
      "step": 9500
    },
    {
      "epoch": 1.8422991893883567,
      "grad_norm": 1.4242560863494873,
      "learning_rate": 2.697126013264554e-08,
      "loss": 0.1799,
      "step": 10000
    },
    {
      "epoch": 1.8422991893883567,
      "eval_loss": 0.9402961730957031,
      "eval_runtime": 74.4464,
      "eval_samples_per_second": 16.213,
      "eval_steps_per_second": 2.028,
      "step": 10000
    },
    {
      "epoch": 1.9344141488577744,
      "grad_norm": 0.06974973529577255,
      "learning_rate": 2.5819823139277818e-08,
      "loss": 0.2114,
      "step": 10500
    },
    {
      "epoch": 1.9344141488577744,
      "eval_loss": 0.9320051074028015,
      "eval_runtime": 76.0983,
      "eval_samples_per_second": 15.861,
      "eval_steps_per_second": 1.984,
      "step": 10500
    },
    {
      "epoch": 2.026529108327192,
      "grad_norm": 0.17525330185890198,
      "learning_rate": 2.4668386145910094e-08,
      "loss": 0.2058,
      "step": 11000
    },
    {
      "epoch": 2.026529108327192,
      "eval_loss": 0.9245836138725281,
      "eval_runtime": 74.5035,
      "eval_samples_per_second": 16.201,
      "eval_steps_per_second": 2.027,
      "step": 11000
    },
    {
      "epoch": 2.1186440677966103,
      "grad_norm": 9.384895324707031,
      "learning_rate": 2.3516949152542374e-08,
      "loss": 0.188,
      "step": 11500
    },
    {
      "epoch": 2.1186440677966103,
      "eval_loss": 0.9178166389465332,
      "eval_runtime": 74.4204,
      "eval_samples_per_second": 16.219,
      "eval_steps_per_second": 2.029,
      "step": 11500
    },
    {
      "epoch": 2.210759027266028,
      "grad_norm": 0.770260214805603,
      "learning_rate": 2.236551215917465e-08,
      "loss": 0.1657,
      "step": 12000
    },
    {
      "epoch": 2.210759027266028,
      "eval_loss": 0.9122205376625061,
      "eval_runtime": 76.1185,
      "eval_samples_per_second": 15.857,
      "eval_steps_per_second": 1.984,
      "step": 12000
    },
    {
      "epoch": 2.3028739867354457,
      "grad_norm": 614.7385864257812,
      "learning_rate": 2.1214075165806926e-08,
      "loss": 0.1948,
      "step": 12500
    },
    {
      "epoch": 2.3028739867354457,
      "eval_loss": 0.90776127576828,
      "eval_runtime": 74.5589,
      "eval_samples_per_second": 16.189,
      "eval_steps_per_second": 2.025,
      "step": 12500
    },
    {
      "epoch": 2.3949889462048635,
      "grad_norm": 0.7871919274330139,
      "learning_rate": 2.0062638172439202e-08,
      "loss": 0.1882,
      "step": 13000
    },
    {
      "epoch": 2.3949889462048635,
      "eval_loss": 0.9029939770698547,
      "eval_runtime": 74.3684,
      "eval_samples_per_second": 16.23,
      "eval_steps_per_second": 2.03,
      "step": 13000
    },
    {
      "epoch": 2.4871039056742816,
      "grad_norm": 1.3432903289794922,
      "learning_rate": 1.891120117907148e-08,
      "loss": 0.1874,
      "step": 13500
    },
    {
      "epoch": 2.4871039056742816,
      "eval_loss": 0.8971672654151917,
      "eval_runtime": 76.1444,
      "eval_samples_per_second": 15.851,
      "eval_steps_per_second": 1.983,
      "step": 13500
    },
    {
      "epoch": 2.5792188651436994,
      "grad_norm": 343.2808532714844,
      "learning_rate": 1.7759764185703757e-08,
      "loss": 0.1423,
      "step": 14000
    },
    {
      "epoch": 2.5792188651436994,
      "eval_loss": 0.8923307657241821,
      "eval_runtime": 74.6823,
      "eval_samples_per_second": 16.162,
      "eval_steps_per_second": 2.022,
      "step": 14000
    },
    {
      "epoch": 2.671333824613117,
      "grad_norm": 0.01757567748427391,
      "learning_rate": 1.6608327192336037e-08,
      "loss": 0.1424,
      "step": 14500
    },
    {
      "epoch": 2.671333824613117,
      "eval_loss": 0.888923704624176,
      "eval_runtime": 74.5202,
      "eval_samples_per_second": 16.197,
      "eval_steps_per_second": 2.026,
      "step": 14500
    },
    {
      "epoch": 2.7634487840825352,
      "grad_norm": 1.6630617380142212,
      "learning_rate": 1.5456890198968313e-08,
      "loss": 0.18,
      "step": 15000
    },
    {
      "epoch": 2.7634487840825352,
      "eval_loss": 0.885770320892334,
      "eval_runtime": 76.123,
      "eval_samples_per_second": 15.856,
      "eval_steps_per_second": 1.984,
      "step": 15000
    },
    {
      "epoch": 2.855563743551953,
      "grad_norm": 0.0005675083957612514,
      "learning_rate": 1.430545320560059e-08,
      "loss": 0.1699,
      "step": 15500
    },
    {
      "epoch": 2.855563743551953,
      "eval_loss": 0.8834386467933655,
      "eval_runtime": 74.4405,
      "eval_samples_per_second": 16.214,
      "eval_steps_per_second": 2.028,
      "step": 15500
    },
    {
      "epoch": 2.9476787030213707,
      "grad_norm": 232.96043395996094,
      "learning_rate": 1.3154016212232866e-08,
      "loss": 0.1711,
      "step": 16000
    },
    {
      "epoch": 2.9476787030213707,
      "eval_loss": 0.8807593584060669,
      "eval_runtime": 74.4453,
      "eval_samples_per_second": 16.213,
      "eval_steps_per_second": 2.028,
      "step": 16000
    },
    {
      "epoch": 3.0397936624907884,
      "grad_norm": 5.004946069675498e-05,
      "learning_rate": 1.2002579218865142e-08,
      "loss": 0.1529,
      "step": 16500
    },
    {
      "epoch": 3.0397936624907884,
      "eval_loss": 0.878940224647522,
      "eval_runtime": 76.1871,
      "eval_samples_per_second": 15.843,
      "eval_steps_per_second": 1.982,
      "step": 16500
    },
    {
      "epoch": 3.131908621960206,
      "grad_norm": 0.01577543281018734,
      "learning_rate": 1.085114222549742e-08,
      "loss": 0.1715,
      "step": 17000
    },
    {
      "epoch": 3.131908621960206,
      "eval_loss": 0.8767727017402649,
      "eval_runtime": 74.5452,
      "eval_samples_per_second": 16.192,
      "eval_steps_per_second": 2.026,
      "step": 17000
    },
    {
      "epoch": 3.2240235814296243,
      "grad_norm": 70.45864868164062,
      "learning_rate": 9.699705232129698e-09,
      "loss": 0.1556,
      "step": 17500
    },
    {
      "epoch": 3.2240235814296243,
      "eval_loss": 0.8749921321868896,
      "eval_runtime": 74.5176,
      "eval_samples_per_second": 16.198,
      "eval_steps_per_second": 2.026,
      "step": 17500
    }
  ],
  "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": false
      },
      "attributes": {}
    }
  },
  "total_flos": 6296490663752520.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}