{
  "best_metric": 0.51348477602005,
  "best_model_checkpoint": "./training_output/swin-base-patch4-window7-224_11092024\\checkpoint-8000",
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 8000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.25,
      "grad_norm": 4.37050724029541,
      "learning_rate": 0.00029624999999999996,
      "loss": 1.6781,
      "step": 100
    },
    {
      "epoch": 0.5,
      "grad_norm": 3.8442654609680176,
      "learning_rate": 0.00029249999999999995,
      "loss": 1.2684,
      "step": 200
    },
    {
      "epoch": 0.75,
      "grad_norm": 3.500859498977661,
      "learning_rate": 0.00028875,
      "loss": 1.0998,
      "step": 300
    },
    {
      "epoch": 1.0,
      "grad_norm": 5.394394874572754,
      "learning_rate": 0.000285,
      "loss": 1.0366,
      "step": 400
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.72,
      "eval_loss": 0.9471067190170288,
      "eval_runtime": 191.6064,
      "eval_samples_per_second": 4.175,
      "eval_steps_per_second": 0.522,
      "step": 400
    },
    {
      "epoch": 1.25,
      "grad_norm": 5.127144813537598,
      "learning_rate": 0.00028125,
      "loss": 0.9272,
      "step": 500
    },
    {
      "epoch": 1.5,
      "grad_norm": 3.240920305252075,
      "learning_rate": 0.00027749999999999997,
      "loss": 0.8831,
      "step": 600
    },
    {
      "epoch": 1.75,
      "grad_norm": 2.9190714359283447,
      "learning_rate": 0.00027374999999999996,
      "loss": 0.8501,
      "step": 700
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.049932956695557,
      "learning_rate": 0.00027,
      "loss": 0.8257,
      "step": 800
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.75375,
      "eval_loss": 0.7888995409011841,
      "eval_runtime": 190.7471,
      "eval_samples_per_second": 4.194,
      "eval_steps_per_second": 0.524,
      "step": 800
    },
    {
      "epoch": 2.25,
      "grad_norm": 2.773448944091797,
      "learning_rate": 0.00026624999999999994,
      "loss": 0.804,
      "step": 900
    },
    {
      "epoch": 2.5,
      "grad_norm": 4.934525489807129,
      "learning_rate": 0.0002625,
      "loss": 0.7622,
      "step": 1000
    },
    {
      "epoch": 2.75,
      "grad_norm": 2.755113363265991,
      "learning_rate": 0.00025875,
      "loss": 0.7747,
      "step": 1100
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.5829150676727295,
      "learning_rate": 0.00025499999999999996,
      "loss": 0.7119,
      "step": 1200
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7775,
      "eval_loss": 0.723233699798584,
      "eval_runtime": 192.2538,
      "eval_samples_per_second": 4.161,
      "eval_steps_per_second": 0.52,
      "step": 1200
    },
    {
      "epoch": 3.25,
      "grad_norm": 3.0522494316101074,
      "learning_rate": 0.00025125,
      "loss": 0.7059,
      "step": 1300
    },
    {
      "epoch": 3.5,
      "grad_norm": 2.336416482925415,
      "learning_rate": 0.00024749999999999994,
      "loss": 0.6803,
      "step": 1400
    },
    {
      "epoch": 3.75,
      "grad_norm": 2.500382900238037,
      "learning_rate": 0.00024375,
      "loss": 0.6944,
      "step": 1500
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.7047812938690186,
      "learning_rate": 0.00023999999999999998,
      "loss": 0.6969,
      "step": 1600
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.78375,
      "eval_loss": 0.673926830291748,
      "eval_runtime": 188.8251,
      "eval_samples_per_second": 4.237,
      "eval_steps_per_second": 0.53,
      "step": 1600
    },
    {
      "epoch": 4.25,
      "grad_norm": 1.99785315990448,
      "learning_rate": 0.00023624999999999997,
      "loss": 0.6529,
      "step": 1700
    },
    {
      "epoch": 4.5,
      "grad_norm": 2.6651065349578857,
      "learning_rate": 0.00023249999999999999,
      "loss": 0.6514,
      "step": 1800
    },
    {
      "epoch": 4.75,
      "grad_norm": 2.7539191246032715,
      "learning_rate": 0.00022874999999999998,
      "loss": 0.6553,
      "step": 1900
    },
    {
      "epoch": 5.0,
      "grad_norm": 2.1046321392059326,
      "learning_rate": 0.000225,
      "loss": 0.6487,
      "step": 2000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.78625,
      "eval_loss": 0.6371492743492126,
      "eval_runtime": 187.904,
      "eval_samples_per_second": 4.257,
      "eval_steps_per_second": 0.532,
      "step": 2000
    },
    {
      "epoch": 5.25,
      "grad_norm": 3.1880531311035156,
      "learning_rate": 0.00022124999999999998,
      "loss": 0.6217,
      "step": 2100
    },
    {
      "epoch": 5.5,
      "grad_norm": 2.7244873046875,
      "learning_rate": 0.00021749999999999997,
      "loss": 0.6168,
      "step": 2200
    },
    {
      "epoch": 5.75,
      "grad_norm": 3.3323047161102295,
      "learning_rate": 0.00021375,
      "loss": 0.6135,
      "step": 2300
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.5545437335968018,
      "learning_rate": 0.00020999999999999998,
      "loss": 0.5956,
      "step": 2400
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.78875,
      "eval_loss": 0.6197744607925415,
      "eval_runtime": 190.0271,
      "eval_samples_per_second": 4.21,
      "eval_steps_per_second": 0.526,
      "step": 2400
    },
    {
      "epoch": 6.25,
      "grad_norm": 2.105663299560547,
      "learning_rate": 0.00020624999999999997,
      "loss": 0.5865,
      "step": 2500
    },
    {
      "epoch": 6.5,
      "grad_norm": 3.24285626411438,
      "learning_rate": 0.0002025,
      "loss": 0.5933,
      "step": 2600
    },
    {
      "epoch": 6.75,
      "grad_norm": 2.4483067989349365,
      "learning_rate": 0.00019874999999999998,
      "loss": 0.6034,
      "step": 2700
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.4036154747009277,
      "learning_rate": 0.000195,
      "loss": 0.5604,
      "step": 2800
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8025,
      "eval_loss": 0.5941271781921387,
      "eval_runtime": 188.5123,
      "eval_samples_per_second": 4.244,
      "eval_steps_per_second": 0.53,
      "step": 2800
    },
    {
      "epoch": 7.25,
      "grad_norm": 2.846735954284668,
      "learning_rate": 0.00019124999999999996,
      "loss": 0.5863,
      "step": 2900
    },
    {
      "epoch": 7.5,
      "grad_norm": 2.242002248764038,
      "learning_rate": 0.00018749999999999998,
      "loss": 0.5762,
      "step": 3000
    },
    {
      "epoch": 7.75,
      "grad_norm": 3.0385332107543945,
      "learning_rate": 0.00018375,
      "loss": 0.5552,
      "step": 3100
    },
    {
      "epoch": 8.0,
      "grad_norm": 4.019121170043945,
      "learning_rate": 0.00017999999999999998,
      "loss": 0.5732,
      "step": 3200
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.795,
      "eval_loss": 0.5866623520851135,
      "eval_runtime": 189.5806,
      "eval_samples_per_second": 4.22,
      "eval_steps_per_second": 0.527,
      "step": 3200
    },
    {
      "epoch": 8.25,
      "grad_norm": 2.7213943004608154,
      "learning_rate": 0.00017625,
      "loss": 0.5913,
      "step": 3300
    },
    {
      "epoch": 8.5,
      "grad_norm": 2.490701675415039,
      "learning_rate": 0.00017249999999999996,
      "loss": 0.5469,
      "step": 3400
    },
    {
      "epoch": 8.75,
      "grad_norm": 2.9301528930664062,
      "learning_rate": 0.00016874999999999998,
      "loss": 0.5188,
      "step": 3500
    },
    {
      "epoch": 9.0,
      "grad_norm": 3.5669283866882324,
      "learning_rate": 0.000165,
      "loss": 0.5578,
      "step": 3600
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.8025,
      "eval_loss": 0.570527195930481,
      "eval_runtime": 187.5953,
      "eval_samples_per_second": 4.265,
      "eval_steps_per_second": 0.533,
      "step": 3600
    },
    {
      "epoch": 9.25,
      "grad_norm": 2.9957895278930664,
      "learning_rate": 0.00016125,
      "loss": 0.532,
      "step": 3700
    },
    {
      "epoch": 9.5,
      "grad_norm": 3.6296515464782715,
      "learning_rate": 0.00015749999999999998,
      "loss": 0.5359,
      "step": 3800
    },
    {
      "epoch": 9.75,
      "grad_norm": 2.170017957687378,
      "learning_rate": 0.00015374999999999997,
      "loss": 0.5334,
      "step": 3900
    },
    {
      "epoch": 10.0,
      "grad_norm": 2.3260116577148438,
      "learning_rate": 0.00015,
      "loss": 0.5449,
      "step": 4000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.81125,
      "eval_loss": 0.5574603080749512,
      "eval_runtime": 186.6415,
      "eval_samples_per_second": 4.286,
      "eval_steps_per_second": 0.536,
      "step": 4000
    },
    {
      "epoch": 10.25,
      "grad_norm": 3.2211897373199463,
      "learning_rate": 0.00014624999999999998,
      "loss": 0.5333,
      "step": 4100
    },
    {
      "epoch": 10.5,
      "grad_norm": 1.6050548553466797,
      "learning_rate": 0.0001425,
      "loss": 0.5206,
      "step": 4200
    },
    {
      "epoch": 10.75,
      "grad_norm": 3.3255255222320557,
      "learning_rate": 0.00013874999999999998,
      "loss": 0.4932,
      "step": 4300
    },
    {
      "epoch": 11.0,
      "grad_norm": 2.705442190170288,
      "learning_rate": 0.000135,
      "loss": 0.5419,
      "step": 4400
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.82125,
      "eval_loss": 0.5505005121231079,
      "eval_runtime": 186.7804,
      "eval_samples_per_second": 4.283,
      "eval_steps_per_second": 0.535,
      "step": 4400
    },
    {
      "epoch": 11.25,
      "grad_norm": 2.7391810417175293,
      "learning_rate": 0.00013125,
      "loss": 0.4945,
      "step": 4500
    },
    {
      "epoch": 11.5,
      "grad_norm": 2.746691942214966,
      "learning_rate": 0.00012749999999999998,
      "loss": 0.5196,
      "step": 4600
    },
    {
      "epoch": 11.75,
      "grad_norm": 3.134870767593384,
      "learning_rate": 0.00012374999999999997,
      "loss": 0.548,
      "step": 4700
    },
    {
      "epoch": 12.0,
      "grad_norm": 3.3967835903167725,
      "learning_rate": 0.00011999999999999999,
      "loss": 0.5086,
      "step": 4800
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.82125,
      "eval_loss": 0.5385071039199829,
      "eval_runtime": 187.7731,
      "eval_samples_per_second": 4.26,
      "eval_steps_per_second": 0.533,
      "step": 4800
    },
    {
      "epoch": 12.25,
      "grad_norm": 2.0934133529663086,
      "learning_rate": 0.00011624999999999999,
      "loss": 0.5059,
      "step": 4900
    },
    {
      "epoch": 12.5,
      "grad_norm": 3.039590835571289,
      "learning_rate": 0.0001125,
      "loss": 0.5043,
      "step": 5000
    },
    {
      "epoch": 12.75,
      "grad_norm": 2.2998437881469727,
      "learning_rate": 0.00010874999999999999,
      "loss": 0.5084,
      "step": 5100
    },
    {
      "epoch": 13.0,
      "grad_norm": 1.5604509115219116,
      "learning_rate": 0.00010499999999999999,
      "loss": 0.4929,
      "step": 5200
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.82125,
      "eval_loss": 0.5340179204940796,
      "eval_runtime": 189.2687,
      "eval_samples_per_second": 4.227,
      "eval_steps_per_second": 0.528,
      "step": 5200
    },
    {
      "epoch": 13.25,
      "grad_norm": 4.638809680938721,
      "learning_rate": 0.00010125,
      "loss": 0.5198,
      "step": 5300
    },
    {
      "epoch": 13.5,
      "grad_norm": 2.361114263534546,
      "learning_rate": 9.75e-05,
      "loss": 0.5016,
      "step": 5400
    },
    {
      "epoch": 13.75,
      "grad_norm": 1.9207624197006226,
      "learning_rate": 9.374999999999999e-05,
      "loss": 0.5064,
      "step": 5500
    },
    {
      "epoch": 14.0,
      "grad_norm": 2.6521153450012207,
      "learning_rate": 8.999999999999999e-05,
      "loss": 0.4701,
      "step": 5600
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.81875,
      "eval_loss": 0.5296584963798523,
      "eval_runtime": 188.8651,
      "eval_samples_per_second": 4.236,
      "eval_steps_per_second": 0.529,
      "step": 5600
    },
    {
      "epoch": 14.25,
      "grad_norm": 1.9743889570236206,
      "learning_rate": 8.624999999999998e-05,
      "loss": 0.5078,
      "step": 5700
    },
    {
      "epoch": 14.5,
      "grad_norm": 2.232062578201294,
      "learning_rate": 8.25e-05,
      "loss": 0.4723,
      "step": 5800
    },
    {
      "epoch": 14.75,
      "grad_norm": 3.261901617050171,
      "learning_rate": 7.874999999999999e-05,
      "loss": 0.5011,
      "step": 5900
    },
    {
      "epoch": 15.0,
      "grad_norm": 2.4302258491516113,
      "learning_rate": 7.5e-05,
      "loss": 0.4803,
      "step": 6000
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.8225,
      "eval_loss": 0.5240346789360046,
      "eval_runtime": 189.3391,
      "eval_samples_per_second": 4.225,
      "eval_steps_per_second": 0.528,
      "step": 6000
    },
    {
      "epoch": 15.25,
      "grad_norm": 2.341607093811035,
      "learning_rate": 7.125e-05,
      "loss": 0.4791,
      "step": 6100
    },
    {
      "epoch": 15.5,
      "grad_norm": 2.3899717330932617,
      "learning_rate": 6.75e-05,
      "loss": 0.4872,
      "step": 6200
    },
    {
      "epoch": 15.75,
      "grad_norm": 2.253009557723999,
      "learning_rate": 6.374999999999999e-05,
      "loss": 0.5044,
      "step": 6300
    },
    {
      "epoch": 16.0,
      "grad_norm": 3.2920048236846924,
      "learning_rate": 5.9999999999999995e-05,
      "loss": 0.4988,
      "step": 6400
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.83,
      "eval_loss": 0.5196738243103027,
      "eval_runtime": 189.4702,
      "eval_samples_per_second": 4.222,
      "eval_steps_per_second": 0.528,
      "step": 6400
    },
    {
      "epoch": 16.25,
      "grad_norm": 4.486973762512207,
      "learning_rate": 5.625e-05,
      "loss": 0.4831,
      "step": 6500
    },
    {
      "epoch": 16.5,
      "grad_norm": 1.689079999923706,
      "learning_rate": 5.2499999999999995e-05,
      "loss": 0.4961,
      "step": 6600
    },
    {
      "epoch": 16.75,
      "grad_norm": 2.2517282962799072,
      "learning_rate": 4.875e-05,
      "loss": 0.489,
      "step": 6700
    },
    {
      "epoch": 17.0,
      "grad_norm": 2.318455219268799,
      "learning_rate": 4.4999999999999996e-05,
      "loss": 0.4842,
      "step": 6800
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.83125,
      "eval_loss": 0.5164589881896973,
      "eval_runtime": 188.3077,
      "eval_samples_per_second": 4.248,
      "eval_steps_per_second": 0.531,
      "step": 6800
    },
    {
      "epoch": 17.25,
      "grad_norm": 3.113399028778076,
      "learning_rate": 4.125e-05,
      "loss": 0.4676,
      "step": 6900
    },
    {
      "epoch": 17.5,
      "grad_norm": 2.0549120903015137,
      "learning_rate": 3.75e-05,
      "loss": 0.4679,
      "step": 7000
    },
    {
      "epoch": 17.75,
      "grad_norm": 2.0566205978393555,
      "learning_rate": 3.375e-05,
      "loss": 0.4988,
      "step": 7100
    },
    {
      "epoch": 18.0,
      "grad_norm": 2.392482280731201,
      "learning_rate": 2.9999999999999997e-05,
      "loss": 0.4917,
      "step": 7200
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.83125,
      "eval_loss": 0.5148141384124756,
      "eval_runtime": 188.4138,
      "eval_samples_per_second": 4.246,
      "eval_steps_per_second": 0.531,
      "step": 7200
    },
    {
      "epoch": 18.25,
      "grad_norm": 1.7607569694519043,
      "learning_rate": 2.6249999999999998e-05,
      "loss": 0.475,
      "step": 7300
    },
    {
      "epoch": 18.5,
      "grad_norm": 2.870453357696533,
      "learning_rate": 2.2499999999999998e-05,
      "loss": 0.469,
      "step": 7400
    },
    {
      "epoch": 18.75,
      "grad_norm": 2.928583860397339,
      "learning_rate": 1.875e-05,
      "loss": 0.4691,
      "step": 7500
    },
    {
      "epoch": 19.0,
      "grad_norm": 3.0488061904907227,
      "learning_rate": 1.4999999999999999e-05,
      "loss": 0.4734,
      "step": 7600
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.8325,
      "eval_loss": 0.5140394568443298,
      "eval_runtime": 187.2755,
      "eval_samples_per_second": 4.272,
      "eval_steps_per_second": 0.534,
      "step": 7600
    },
    {
      "epoch": 19.25,
      "grad_norm": 3.021597146987915,
      "learning_rate": 1.1249999999999999e-05,
      "loss": 0.5122,
      "step": 7700
    },
    {
      "epoch": 19.5,
      "grad_norm": 1.8236939907073975,
      "learning_rate": 7.499999999999999e-06,
      "loss": 0.4582,
      "step": 7800
    },
    {
      "epoch": 19.75,
      "grad_norm": 1.9407800436019897,
      "learning_rate": 3.7499999999999997e-06,
      "loss": 0.4578,
      "step": 7900
    },
    {
      "epoch": 20.0,
      "grad_norm": 2.3664169311523438,
      "learning_rate": 0.0,
      "loss": 0.4714,
      "step": 8000
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.83375,
      "eval_loss": 0.51348477602005,
      "eval_runtime": 192.831,
      "eval_samples_per_second": 4.149,
      "eval_steps_per_second": 0.519,
      "step": 8000
    },
    {
      "epoch": 20.0,
      "step": 8000,
      "total_flos": 1.0028942094237696e+19,
      "train_loss": 0.6026930379867553,
      "train_runtime": 35487.7868,
      "train_samples_per_second": 3.607,
      "train_steps_per_second": 0.225
    }
  ],
  "logging_steps": 100,
  "max_steps": 8000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "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": 1.0028942094237696e+19,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}