checkpoint-277/trainer_state.json

{
  "best_metric": 0.8839941620826721,
  "best_model_checkpoint": "outputs/checkpoint-185",
  "epoch": 5.98110661268556,
  "eval_steps": 500,
  "global_step": 277,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.4318488529014845,
      "grad_norm": 0.6420097947120667,
      "learning_rate": 4e-05,
      "loss": 1.9916,
      "step": 20
    },
    {
      "epoch": 0.863697705802969,
      "grad_norm": 0.48513880372047424,
      "learning_rate": 8e-05,
      "loss": 1.8266,
      "step": 40
    },
    {
      "epoch": 0.9932523616734144,
      "eval_loss": 1.5141435861587524,
      "eval_runtime": 12.7002,
      "eval_samples_per_second": 29.291,
      "eval_steps_per_second": 3.701,
      "step": 46
    },
    {
      "epoch": 1.2955465587044535,
      "grad_norm": 0.711148202419281,
      "learning_rate": 0.00012,
      "loss": 1.7047,
      "step": 60
    },
    {
      "epoch": 1.7273954116059378,
      "grad_norm": 0.805604875087738,
      "learning_rate": 0.00016,
      "loss": 1.481,
      "step": 80
    },
    {
      "epoch": 1.9865047233468287,
      "eval_loss": 1.1201857328414917,
      "eval_runtime": 12.66,
      "eval_samples_per_second": 29.384,
      "eval_steps_per_second": 3.712,
      "step": 92
    },
    {
      "epoch": 2.1592442645074224,
      "grad_norm": 1.1893383264541626,
      "learning_rate": 0.0002,
      "loss": 1.362,
      "step": 100
    },
    {
      "epoch": 2.591093117408907,
      "grad_norm": 1.204134464263916,
      "learning_rate": 0.00019943348002101371,
      "loss": 1.1848,
      "step": 120
    },
    {
      "epoch": 2.979757085020243,
      "eval_loss": 0.9436031579971313,
      "eval_runtime": 12.6699,
      "eval_samples_per_second": 29.361,
      "eval_steps_per_second": 3.71,
      "step": 138
    },
    {
      "epoch": 3.0229419703103915,
      "grad_norm": 1.0650242567062378,
      "learning_rate": 0.00019774033898178667,
      "loss": 1.154,
      "step": 140
    },
    {
      "epoch": 3.454790823211876,
      "grad_norm": 1.2103493213653564,
      "learning_rate": 0.00019493976084683813,
      "loss": 0.9441,
      "step": 160
    },
    {
      "epoch": 3.8866396761133606,
      "grad_norm": 1.035901427268982,
      "learning_rate": 0.00019106347728549135,
      "loss": 0.9405,
      "step": 180
    },
    {
      "epoch": 3.9946018893387314,
      "eval_loss": 0.8839941620826721,
      "eval_runtime": 12.6739,
      "eval_samples_per_second": 29.352,
      "eval_steps_per_second": 3.708,
      "step": 185
    },
    {
      "epoch": 4.318488529014845,
      "grad_norm": 1.2742565870285034,
      "learning_rate": 0.0001861554081393806,
      "loss": 0.7772,
      "step": 200
    },
    {
      "epoch": 4.75033738191633,
      "grad_norm": 1.2412225008010864,
      "learning_rate": 0.00018027116379309638,
      "loss": 0.8008,
      "step": 220
    },
    {
      "epoch": 4.987854251012146,
      "eval_loss": 0.9072163701057434,
      "eval_runtime": 12.6529,
      "eval_samples_per_second": 29.4,
      "eval_steps_per_second": 3.715,
      "step": 231
    },
    {
      "epoch": 5.182186234817814,
      "grad_norm": 1.1546653509140015,
      "learning_rate": 0.00017347741508630672,
      "loss": 0.7324,
      "step": 240
    },
    {
      "epoch": 5.614035087719298,
      "grad_norm": 1.1490098237991333,
      "learning_rate": 0.00016585113790650388,
      "loss": 0.6233,
      "step": 260
    },
    {
      "epoch": 5.98110661268556,
      "eval_loss": 0.9705129861831665,
      "eval_runtime": 12.6584,
      "eval_samples_per_second": 29.388,
      "eval_steps_per_second": 3.713,
      "step": 277
    }
  ],
  "logging_steps": 20,
  "max_steps": 690,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "save_steps": 500,
  "total_flos": 2.134024227305472e+16,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}