{
  "best_metric": 3.3192038536071777,
  "best_model_checkpoint": "/users/zyong2/data/zyong2/bigscience/data/processed/024/bloom-350m_de_sft_100000samples_-1vocab_original-frozen/checkpoint-25000",
  "epoch": 2.758772897815052,
  "global_step": 25000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.28,
      "l1_reg_loss": 0.0,
      "learning_rate": 9e-05,
      "loss": 3.8882,
      "step": 2500
    },
    {
      "epoch": 0.55,
      "l1_reg_loss": 0.0,
      "learning_rate": 8e-05,
      "loss": 3.6037,
      "step": 5000
    },
    {
      "epoch": 0.55,
      "eval_loss": 3.5761282444000244,
      "eval_runtime": 397.3988,
      "eval_samples_per_second": 9.592,
      "eval_steps_per_second": 4.796,
      "step": 5000
    },
    {
      "epoch": 0.83,
      "l1_reg_loss": 0.0001,
      "learning_rate": 7e-05,
      "loss": 3.4802,
      "step": 7500
    },
    {
      "epoch": 1.1,
      "l1_reg_loss": 0.0001,
      "learning_rate": 6e-05,
      "loss": 3.4086,
      "step": 10000
    },
    {
      "epoch": 1.1,
      "eval_loss": 3.4271042346954346,
      "eval_runtime": 398.9002,
      "eval_samples_per_second": 9.556,
      "eval_steps_per_second": 4.778,
      "step": 10000
    },
    {
      "epoch": 1.38,
      "l1_reg_loss": 0.0001,
      "learning_rate": 5e-05,
      "loss": 3.343,
      "step": 12500
    },
    {
      "epoch": 1.66,
      "l1_reg_loss": 0.0001,
      "learning_rate": 4e-05,
      "loss": 3.309,
      "step": 15000
    },
    {
      "epoch": 1.66,
      "eval_loss": 3.3605945110321045,
      "eval_runtime": 398.3343,
      "eval_samples_per_second": 9.57,
      "eval_steps_per_second": 4.785,
      "step": 15000
    },
    {
      "epoch": 1.93,
      "l1_reg_loss": 0.0001,
      "learning_rate": 3e-05,
      "loss": 3.2949,
      "step": 17500
    },
    {
      "epoch": 2.21,
      "l1_reg_loss": 0.0001,
      "learning_rate": 2e-05,
      "loss": 3.2684,
      "step": 20000
    },
    {
      "epoch": 2.21,
      "eval_loss": 3.3289120197296143,
      "eval_runtime": 398.231,
      "eval_samples_per_second": 9.572,
      "eval_steps_per_second": 4.786,
      "step": 20000
    },
    {
      "epoch": 2.48,
      "l1_reg_loss": 0.0001,
      "learning_rate": 1e-05,
      "loss": 3.2569,
      "step": 22500
    },
    {
      "epoch": 2.76,
      "l1_reg_loss": 0.0001,
      "learning_rate": 0.0,
      "loss": 3.2477,
      "step": 25000
    },
    {
      "epoch": 2.76,
      "eval_loss": 3.3192038536071777,
      "eval_runtime": 397.9798,
      "eval_samples_per_second": 9.578,
      "eval_steps_per_second": 4.789,
      "step": 25000
    },
    {
      "epoch": 2.76,
      "step": 25000,
      "total_flos": 3.7147907956565606e+17,
      "train_loss": 3.41004751953125,
      "train_runtime": 60598.6579,
      "train_samples_per_second": 3.3,
      "train_steps_per_second": 0.413
    }
  ],
  "max_steps": 25000,
  "num_train_epochs": 3,
  "total_flos": 3.7147907956565606e+17,
  "trial_name": null,
  "trial_params": null
}