| { |
| "parameter_count": 205783040, |
| "parameter_summary": { |
| "total": 205783040, |
| "parts": { |
| "blocks": 102760448, |
| "lm_head": 51511296, |
| "other": 51511296 |
| } |
| }, |
| "n_layer": 8, |
| "n_head": 8, |
| "n_embd": 1024, |
| "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": 0.04, |
| "embedding_lr": 0.01810193359836, |
| "unembedding_lr": 0.0045254833996, |
| "lr_multiplier": 1.0, |
| "effective_matrix_lr": 0.04, |
| "effective_embedding_lr": 0.01810193359836, |
| "effective_unembedding_lr": 0.0045254833996, |
| "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": 32, |
| "num_epochs": 1, |
| "estimated_num_iterations": 1980, |
| "lr_schedule_steps": 1962, |
| "resolved_lr_schedule_steps": 1962, |
| "max_train_steps": 1962, |
| "resolved_max_train_steps": 1962, |
| "completed_epochs": 0, |
| "seed": 42, |
| "train_token_limit": 1038939586, |
| "val_token_limit": 10000000, |
| "token_limit_cutoff_devices": 8, |
| "active_train_tokens": 1038939586, |
| "packed_train_sequences_per_epoch": 506880, |
| "packed_train_tokens_per_epoch": 1038090240, |
| "active_val_packed_sequences": 4864, |
| "active_val_packed_tokens": 9961472, |
| "val_loss": 3.329581009714227, |
| "final_val_loss": 3.329581009714227, |
| "final_train_loss": 3.3690894610840303, |
| "best_val_loss": 3.329581009714227, |
| "best_val_bpb": 1.0821464569815382, |
| "best_val_epoch": 1.0, |
| "train_val_per_epoch": [ |
| { |
| "epoch": 1, |
| "step": 1962, |
| "train_loss": 3.3690894610840303, |
| "val_loss": 3.329581009714227, |
| "val_bpb": 1.0821464569815382 |
| } |
| ], |
| "core_block_residual_rms": null, |
| "last_train_mfu": 42.49351126377972, |
| "last_train_tok_per_sec": 3275792, |
| "last_train_step_time_seconds": 0.1600492000579834, |
| "total_training_time_seconds": 316.28008460998535, |
| "total_wall_time_seconds": 344.73980259895325, |
| "sequence_len": 2048, |
| "window_pattern": "L" |
| } |
|
|
