{
  "parameter_count": 1378418688,
  "parameter_summary": {
    "total": 1378418688,
    "parts": {
      "blocks": 1146617856,
      "lm_head": 115900416,
      "other": 115900416
    }
  },
  "n_layer": 18,
  "n_head": 18,
  "n_embd": 2304,
  "head_dim": 128,
  "depth_scale_mode": "loop",
  "core_repetitions": 4,
  "core_layer_count": 6,
  "resolved_core_layer_count": 6,
  "prelude_layer_count": 6,
  "resolved_prelude_layer_count": 6,
  "coda_layer_count": 6,
  "resolved_coda_layer_count": 6,
  "effective_depth": 36,
  "expected_effective_depth": 36.0,
  "recurrence_rmsnorm": "recurrence_only",
  "recurrence_emb_alpha": 0.707106781187,
  "decoder_inject": true,
  "core_residual_branch_scale": 1.0,
  "resolved_core_residual_branch_scale": null,
  "core_repetition_schedule": "late",
  "core_repetition_crossover_fraction": 0.87,
  "core_repetition_low": 2,
  "core_repetition_breakpoints": null,
  "lsched_lr_warmup_steps": 40,
  "lsched_lr_change_steps": [
    15098
  ],
  "dep_grow_events": [],
  "target_core_repetitions": 4,
  "final_eval_active_core_repetitions": 4,
  "observed_train_core_repetitions_mean": 2.260097954479977,
  "observed_train_core_repetitions_min": 2,
  "observed_train_core_repetitions_max": 4,
  "observed_train_core_repetitions_histogram": {
    "2": 15098,
    "4": 2257
  },
  "recurrence_truncation_schedule": "none",
  "recurrence_bp_window": 0,
  "matrix_lr": 0.015455184874487526,
  "embedding_lr": 0.004945659159836009,
  "unembedding_lr": 0.0017485545646810576,
  "lr_multiplier": 1.0,
  "effective_matrix_lr": 0.015455184874487526,
  "effective_embedding_lr": 0.004945659159836009,
  "effective_unembedding_lr": 0.0017485545646810576,
  "weight_decay": 0.05,
  "matrix_weight_decay": 0.05,
  "embedding_weight_decay": 0.05,
  "unembedding_weight_decay": 0.05,
  "embedding_wd_skip_norm": 0.0,
  "adam_betas": [
    0.8,
    0.95
  ],
  "adam_eps": 1e-10,
  "lr_schedule": "wsd",
  "warmup_steps": 40,
  "warmdown_ratio": 1.0,
  "output_multiplier": 1.0,
  "residual_branch_multiplier": 0.25,
  "wte_init_std": 0.02,
  "uniform_init_scale": 0.0441941738242,
  "total_batch_size": 524288,
  "device_batch_size": 4,
  "num_epochs": 1,
  "estimated_num_iterations": 17519,
  "lr_schedule_steps": 17355,
  "resolved_lr_schedule_steps": 17355,
  "max_train_steps": 17355,
  "resolved_max_train_steps": 17355,
  "completed_epochs": 0,
  "seed": 42,
  "train_token_limit": 9190008422,
  "val_token_limit": 10000000,
  "token_limit_cutoff_devices": 64,
  "active_train_tokens": 9190008422,
  "packed_train_sequences_per_epoch": 4484864,
  "packed_train_tokens_per_epoch": 9185001472,
  "active_val_packed_sequences": 4864,
  "active_val_packed_tokens": 9961472,
  "val_loss": 2.6445254275673316,
  "final_val_loss": 2.6445254275673316,
  "final_train_loss": 2.6106741496349093,
  "best_val_loss": 2.6445254275673316,
  "best_val_bpb": 0.859490350688765,
  "best_val_epoch": 1.0,
  "train_val_per_epoch": [
    {
      "epoch": 1,
      "step": 17355,
      "train_loss": 2.6106741496349093,
      "val_loss": 2.6445254275673316,
      "val_bpb": 0.859490350688765
    }
  ],
  "core_block_residual_rms": 1.1073907613754272,
  "last_train_mfu": 48.55013989418503,
  "last_train_tok_per_sec": 248232,
  "last_train_step_time_seconds": 2.112081289291382,
  "total_training_time_seconds": 27000.957017183304,
  "total_wall_time_seconds": 27245.46728038788,
  "sequence_len": 2048,
  "window_pattern": "L"
}