{
  "parameter_count": 1843527680,
  "parameter_summary": {
    "total": 1843527680,
    "parts": {
      "blocks": 1585971200,
      "lm_head": 128778240,
      "other": 128778240
    }
  },
  "n_layer": 20,
  "n_head": 20,
  "n_embd": 2560,
  "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": 20,
  "expected_effective_depth": 20.0,
  "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": 0.00692257905530223,
  "embedding_lr": 0.0031328016597119666,
  "unembedding_lr": 0.0007832004149297223,
  "lr_multiplier": 1.0,
  "effective_matrix_lr": 0.00692257905530223,
  "effective_embedding_lr": 0.0031328016597119666,
  "effective_unembedding_lr": 0.0007832004149297223,
  "weight_decay": 0.0707106781187,
  "matrix_weight_decay": 0.0707106781187,
  "embedding_weight_decay": 0.0707106781187,
  "unembedding_weight_decay": 0.0707106781187,
  "embedding_wd_skip_norm": 0.0,
  "adam_betas": [
    0.8,
    0.95
  ],
  "adam_eps": 1e-10,
  "lr_schedule": "wsd",
  "warmup_steps": 40,
  "warmdown_ratio": 0.6,
  "output_multiplier": 0.5,
  "residual_branch_multiplier": 0.25,
  "wte_init_std": 0.00707106781187,
  "uniform_init_scale": 0.0625,
  "total_batch_size": 524288,
  "device_batch_size": 8,
  "num_epochs": 1,
  "estimated_num_iterations": 17748,
  "lr_schedule_steps": 17581,
  "resolved_lr_schedule_steps": 17581,
  "max_train_steps": 17581,
  "resolved_max_train_steps": 17581,
  "completed_epochs": 0,
  "seed": 42,
  "train_token_limit": 9309682401,
  "val_token_limit": 10000000,
  "token_limit_cutoff_devices": 32,
  "active_train_tokens": 9309682401,
  "packed_train_sequences_per_epoch": 4543488,
  "packed_train_tokens_per_epoch": 9305063424,
  "active_val_packed_sequences": 4864,
  "active_val_packed_tokens": 9961472,
  "val_loss": 2.6834234940378288,
  "final_val_loss": 2.6834234940378288,
  "final_train_loss": 2.70216346543036,
  "best_val_loss": 2.6834234940378288,
  "best_val_bpb": 0.8721358614246806,
  "best_val_epoch": 1.0,
  "train_val_per_epoch": [
    {
      "epoch": 1,
      "step": 17581,
      "train_loss": 2.70216346543036,
      "val_loss": 2.6834234940378288,
      "val_bpb": 0.8721358614246806
    }
  ],
  "core_block_residual_rms": 1.6338139772415161,
  "last_train_mfu": 49.79773741120276,
  "last_train_tok_per_sec": 360873,
  "last_train_step_time_seconds": 1.4528310298919678,
  "total_training_time_seconds": 25629.825414419174,
  "total_wall_time_seconds": 25814.953799009323,
  "sequence_len": 2048,
  "window_pattern": "L"
}