{
  "parameter_count": 133890048,
  "parameter_summary": {
    "total": 133890048,
    "parts": {
      "blocks": 56623104,
      "lm_head": 38633472,
      "other": 38633472
    }
  },
  "n_layer": 8,
  "n_head": 6,
  "n_embd": 768,
  "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": 8,
  "expected_effective_depth": 8.0,
  "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": 0.06094039533416783,
  "embedding_lr": 0.009750463253466853,
  "unembedding_lr": 0.0034473093431168786,
  "lr_multiplier": 1.0,
  "effective_matrix_lr": 0.06094039533416783,
  "effective_embedding_lr": 0.009750463253466853,
  "effective_unembedding_lr": 0.0034473093431168786,
  "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": 32,
  "num_epochs": 1,
  "estimated_num_iterations": 1546,
  "lr_schedule_steps": 1532,
  "resolved_lr_schedule_steps": 1532,
  "max_train_steps": 1532,
  "resolved_max_train_steps": 1532,
  "completed_epochs": 0,
  "seed": 42,
  "train_token_limit": 811241308,
  "val_token_limit": 10000000,
  "token_limit_cutoff_devices": 8,
  "active_train_tokens": 811241308,
  "packed_train_sequences_per_epoch": 395776,
  "packed_train_tokens_per_epoch": 810549248,
  "active_val_packed_sequences": 4864,
  "active_val_packed_tokens": 9961472,
  "val_loss": 3.4499592028166117,
  "final_val_loss": 3.4499592028166117,
  "final_train_loss": 3.457156043686349,
  "best_val_loss": 3.4499592028166117,
  "best_val_bpb": 1.1212617648038885,
  "best_val_epoch": 1.0,
  "train_val_per_epoch": [
    {
      "epoch": 1,
      "step": 1532,
      "train_loss": 3.457156043686349,
      "val_loss": 3.4499592028166117,
      "val_bpb": 1.1212617648038885
    }
  ],
  "core_block_residual_rms": null,
  "last_train_mfu": 37.2371384540415,
  "last_train_tok_per_sec": 4553116,
  "last_train_step_time_seconds": 0.11514925956726074,
  "total_training_time_seconds": 176.4720356464386,
  "total_wall_time_seconds": 214.74119806289673,
  "sequence_len": 2048,
  "window_pattern": "L"
}