trainer_state.json

{
  "best_metric": 0.23666666666666666,
  "best_model_checkpoint": "clip-vit-large-patch14-finetuned-clip-vit-large-patch14-mnist_linear_probe/checkpoint-168",
  "epoch": 2.985781990521327,
  "eval_steps": 500,
  "global_step": 168,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.17772511848341233,
      "grad_norm": 2.373176097869873,
      "learning_rate": 2.9411764705882354e-05,
      "loss": 2.4028,
      "step": 10
    },
    {
      "epoch": 0.35545023696682465,
      "grad_norm": 2.2507705688476562,
      "learning_rate": 4.900662251655629e-05,
      "loss": 2.3848,
      "step": 20
    },
    {
      "epoch": 0.533175355450237,
      "grad_norm": 1.5781500339508057,
      "learning_rate": 4.56953642384106e-05,
      "loss": 2.3427,
      "step": 30
    },
    {
      "epoch": 0.7109004739336493,
      "grad_norm": 1.5205079317092896,
      "learning_rate": 4.23841059602649e-05,
      "loss": 2.3118,
      "step": 40
    },
    {
      "epoch": 0.8886255924170616,
      "grad_norm": 1.119712233543396,
      "learning_rate": 3.907284768211921e-05,
      "loss": 2.2886,
      "step": 50
    },
    {
      "epoch": 0.995260663507109,
      "eval_accuracy": 0.157,
      "eval_loss": 2.2661008834838867,
      "eval_runtime": 478.4456,
      "eval_samples_per_second": 12.541,
      "eval_steps_per_second": 0.393,
      "step": 56
    },
    {
      "epoch": 1.066350710900474,
      "grad_norm": 1.0753259658813477,
      "learning_rate": 3.576158940397351e-05,
      "loss": 2.2701,
      "step": 60
    },
    {
      "epoch": 1.2440758293838863,
      "grad_norm": 0.8510423302650452,
      "learning_rate": 3.2450331125827816e-05,
      "loss": 2.2516,
      "step": 70
    },
    {
      "epoch": 1.4218009478672986,
      "grad_norm": 0.9210019111633301,
      "learning_rate": 2.913907284768212e-05,
      "loss": 2.2406,
      "step": 80
    },
    {
      "epoch": 1.599526066350711,
      "grad_norm": 0.9463822245597839,
      "learning_rate": 2.5827814569536424e-05,
      "loss": 2.2312,
      "step": 90
    },
    {
      "epoch": 1.7772511848341233,
      "grad_norm": 0.8657775521278381,
      "learning_rate": 2.2516556291390732e-05,
      "loss": 2.2231,
      "step": 100
    },
    {
      "epoch": 1.9549763033175356,
      "grad_norm": 0.8078551292419434,
      "learning_rate": 1.9205298013245036e-05,
      "loss": 2.2153,
      "step": 110
    },
    {
      "epoch": 1.9905213270142181,
      "eval_accuracy": 0.1945,
      "eval_loss": 2.2003655433654785,
      "eval_runtime": 527.8354,
      "eval_samples_per_second": 11.367,
      "eval_steps_per_second": 0.356,
      "step": 112
    },
    {
      "epoch": 2.132701421800948,
      "grad_norm": 0.7690733075141907,
      "learning_rate": 1.589403973509934e-05,
      "loss": 2.2121,
      "step": 120
    },
    {
      "epoch": 2.31042654028436,
      "grad_norm": 0.6478149890899658,
      "learning_rate": 1.2582781456953644e-05,
      "loss": 2.2047,
      "step": 130
    },
    {
      "epoch": 2.4881516587677726,
      "grad_norm": 0.6131073832511902,
      "learning_rate": 9.271523178807948e-06,
      "loss": 2.2018,
      "step": 140
    },
    {
      "epoch": 2.665876777251185,
      "grad_norm": 0.7310860753059387,
      "learning_rate": 5.9602649006622515e-06,
      "loss": 2.2025,
      "step": 150
    },
    {
      "epoch": 2.843601895734597,
      "grad_norm": 0.6716310381889343,
      "learning_rate": 2.6490066225165563e-06,
      "loss": 2.1981,
      "step": 160
    },
    {
      "epoch": 2.985781990521327,
      "eval_accuracy": 0.23666666666666666,
      "eval_loss": 2.182861566543579,
      "eval_runtime": 73.356,
      "eval_samples_per_second": 81.793,
      "eval_steps_per_second": 2.563,
      "step": 168
    },
    {
      "epoch": 2.985781990521327,
      "step": 168,
      "total_flos": 4.411646023570175e+19,
      "train_loss": 2.258255742845081,
      "train_runtime": 9823.5604,
      "train_samples_per_second": 16.491,
      "train_steps_per_second": 0.017
    }
  ],
  "logging_steps": 10,
  "max_steps": 168,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 4.411646023570175e+19,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}