{
  "parameter_count": 607518720,
  "parameter_summary": {
    "total": 607518720,
    "parts": {
      "blocks": 452984832,
      "lm_head": 77266944,
      "other": 77266944
    }
  },
  "n_layer": 16,
  "n_head": 12,
  "n_embd": 1536,
  "head_dim": 128,
  "depth_scale_mode": "none",
  "core_repetitions": 1,
  "core_layer_count": null,
  "resolved_core_layer_count": null,
  "prelude_layer_count": null,
  "resolved_prelude_layer_count": null,
  "coda_layer_count": null,
  "resolved_coda_layer_count": null,
  "effective_depth": 16,
  "expected_effective_depth": 16.0,
  "recurrence_rmsnorm": "none",
  "recurrence_emb_alpha": 0.0,
  "decoder_inject": false,
  "core_residual_branch_scale": 1.0,
  "resolved_core_residual_branch_scale": null,
  "core_repetition_schedule": "none",
  "core_repetition_crossover_fraction": null,
  "core_repetition_low": 1,
  "core_repetition_breakpoints": null,
  "lsched_lr_warmup_steps": 0,
  "lsched_lr_change_steps": [],
  "dep_grow_events": [],
  "target_core_repetitions": 1,
  "final_eval_active_core_repetitions": null,
  "observed_train_core_repetitions_mean": null,
  "observed_train_core_repetitions_min": null,
  "observed_train_core_repetitions_max": null,
  "observed_train_core_repetitions_histogram": null,
  "recurrence_truncation_schedule": "none",
  "recurrence_bp_window": 0,
  "matrix_lr": 0.021141573630189345,
  "embedding_lr": 0.0033826517808302954,
  "unembedding_lr": 0.0011959480063084232,
  "lr_multiplier": 1.0,
  "effective_matrix_lr": 0.021141573630189345,
  "effective_embedding_lr": 0.0033826517808302954,
  "effective_unembedding_lr": 0.0011959480063084232,
  "weight_decay": 0.1,
  "matrix_weight_decay": 0.1,
  "embedding_weight_decay": 0.1,
  "unembedding_weight_decay": 0.1,
  "embedding_wd_skip_norm": 0.0,
  "adam_betas": [
    0.8,
    0.99
  ],
  "adam_eps": 1e-08,
  "lr_schedule": "wsd",
  "warmup_steps": 5,
  "warmdown_ratio": 0.8,
  "output_multiplier": 1.0,
  "residual_branch_multiplier": 0.5,
  "wte_init_std": 0.005,
  "uniform_init_scale": 0.5,
  "total_batch_size": 524288,
  "device_batch_size": 16,
  "num_epochs": 1,
  "estimated_num_iterations": 7018,
  "lr_schedule_steps": 6952,
  "resolved_lr_schedule_steps": 6952,
  "max_train_steps": 6952,
  "resolved_max_train_steps": 6952,
  "completed_epochs": 0,
  "seed": 42,
  "train_token_limit": 3681298677,
  "val_token_limit": 10000000,
  "token_limit_cutoff_devices": 16,
  "active_train_tokens": 3681298677,
  "packed_train_sequences_per_epoch": 1796608,
  "packed_train_tokens_per_epoch": 3679453184,
  "active_val_packed_sequences": 4864,
  "active_val_packed_tokens": 9961472,
  "val_loss": 2.9291616740979647,
  "final_val_loss": 2.9291616740979647,
  "final_train_loss": 2.9520761289093986,
  "best_val_loss": 2.9291616740979647,
  "best_val_bpb": 0.9519933363650187,
  "best_val_epoch": 1.0,
  "train_val_per_epoch": [
    {
      "epoch": 1,
      "step": 6952,
      "train_loss": 2.9520761289093986,
      "val_loss": 2.9291616740979647,
      "val_bpb": 0.9519933363650187
    }
  ],
  "core_block_residual_rms": 3.235067129135132,
  "last_train_mfu": 46.57236942525834,
  "last_train_tok_per_sec": 1057743,
  "last_train_step_time_seconds": 0.49566650390625,
  "total_training_time_seconds": 3468.908513069153,
  "total_wall_time_seconds": 3576.850750207901,
  "sequence_len": 2048,
  "window_pattern": "L"
}