{ "data": { "bank_fee_probability": 0.08, "candidate_amount_absolute_tolerance": 1.0, "candidate_amount_relative_tolerance": 0.015, "candidate_date_window": 4, "date_horizon_days": 31, "date_shift_days": 3, "description_typo_probability": 0.12, "duplicate_probability": 0.1, "max_amount": 25000.0, "max_bank_rows": 24, "max_events": 16, "max_gl_rows": 24, "min_amount": 5.0, "min_events": 8, "missing_bank_probability": 0.08, "missing_gl_probability": 0.1, "num_accounts": 24, "num_counterparties": 96, "num_currencies": 3, "num_descriptions": 256, "num_entities": 4, "num_references": 8192, "num_samples": 4096, "reference_loss_probability": 0.12, "reference_typo_probability": 0.1, "reversal_probability": 0.08, "seed": 17, "timing_exception_days": 1 }, "loss": { "exception_weight": 0.35, "match_weight": 1.0, "one_to_one_weight": 0.05, "unmatched_class_weight": 0.75 }, "mindweave": { "provenance": { "bank_dataset_id": "mindweave/bank-transactions-us", "bank_dataset_url": "https://huggingface.co/datasets/mindweave/bank-transactions-us", "bank_fee_label_rule": "zero provenance target candidates, blank source_id, source_module=banking, transaction_type=fee, negative amount, and a description containing both fee and a bank/account/service/monthly/maintenance/overdraft/wire term", "bank_revision": "290e657395a1ce25837642b8309ca9a741f85303", "commercial_use_allowed": false, "file_sha256": { "bank_accounts.csv": "bc3a4d01e9924745ae0f8371acfd7caf36a7810ecdc5c24d6e6bb791b4a9ca07", "bank_transactions.csv": "1b20b074ecd254ba0e6056f2e2f196ebff7273aeabc54e389958cd7bbd048604", "chart_of_accounts.csv": "156e10c1bdcd89f1d05da5ece223866ee2ad9d070868b0b484b10c5f75c1a6a8", "journal_entries.csv": "5f8491158bf8c02f7b67f433e970ac5b4acee624eb3eddeede9ffea170adc819", "journal_entry_lines.csv": "456b549d1ca2a55d407a033b3a0dfa0033d2376252406cc05371f9812a5c40a9" }, "label_method": "unique (company_id, source_module, source_id) journal-entry provenance joined to the bank account's mapped cash GL line", "ledger_dataset_id": "mindweave/accounting-ledger-us", "ledger_dataset_url": "https://huggingface.co/datasets/mindweave/accounting-ledger-us", "ledger_revision": "b2f2d0890353438b4473f2c3b2d2b5e15a5132a0", "license": "CC-BY-NC-4.0", "unresolved_policy": "retain the bank row, use match_target=-100, and classify REVIEW; never infer unmatched solely from a missing provenance join" }, "test": { "bank_rows": 702, "cases": 30, "inferred_bank_fees": 3, "periods": [ "2025-10", "2025-11", "2025-12" ], "provenance_matches": 699, "unresolved_rows": 0 }, "train": { "bank_rows": 2021, "cases": 92, "inferred_bank_fees": 18, "periods": [ "2024-01", "2024-02", "2024-03", "2024-04", "2024-05", "2024-06", "2024-07", "2024-08", "2024-09", "2024-10", "2024-11", "2024-12", "2025-01", "2025-02", "2025-03", "2025-04", "2025-05", "2025-06" ], "provenance_matches": 2003, "unresolved_rows": 0 }, "validation": { "bank_rows": 509, "cases": 23, "inferred_bank_fees": 3, "periods": [ "2025-07", "2025-08", "2025-09" ], "provenance_matches": 506, "unresolved_rows": 0 } }, "model": { "H_cycles": 2, "H_layers": 1, "L_cycles": 2, "L_layers": 1, "attention_dropout": 0.0, "category_cardinalities": [ 4, 5, 25, 97, 8193, 257 ], "continuous_dim": 3, "expansion": 4.0, "hidden_size": 128, "init_std": null, "mask_logit_value": -10000.0, "max_bank_rows": 24, "max_gl_rows": 24, "num_exception_types": 6, "num_heads": 4, "rms_norm_eps": 1e-05 }, "synthetic_retention": { "bank_fee_probability": 0.08, "candidate_amount_absolute_tolerance": 1.0, "candidate_amount_relative_tolerance": 0.015, "candidate_date_window": 4, "date_horizon_days": 31, "date_shift_days": 3, "description_typo_probability": 0.12, "duplicate_probability": 0.1, "max_amount": 25000.0, "max_bank_rows": 24, "max_events": 16, "max_gl_rows": 24, "min_amount": 5.0, "min_events": 8, "missing_bank_probability": 0.08, "missing_gl_probability": 0.1, "num_accounts": 24, "num_counterparties": 96, "num_currencies": 3, "num_descriptions": 256, "num_entities": 4, "num_references": 8192, "num_samples": 512, "reference_loss_probability": 0.12, "reference_typo_probability": 0.1, "reversal_probability": 0.08, "seed": 8000074, "timing_exception_days": 1 }, "synthetic_training": { "bank_fee_probability": 0.08, "candidate_amount_absolute_tolerance": 1.0, "candidate_amount_relative_tolerance": 0.015, "candidate_date_window": 4, "date_horizon_days": 31, "date_shift_days": 3, "description_typo_probability": 0.12, "duplicate_probability": 0.1, "max_amount": 25000.0, "max_bank_rows": 24, "max_events": 16, "max_gl_rows": 24, "min_amount": 5.0, "min_events": 8, "missing_bank_probability": 0.08, "missing_gl_probability": 0.1, "num_accounts": 24, "num_counterparties": 96, "num_currencies": 3, "num_descriptions": 256, "num_entities": 4, "num_references": 8192, "num_samples": 512, "reference_loss_probability": 0.12, "reference_typo_probability": 0.1, "reversal_probability": 0.08, "seed": 7000060, "timing_exception_days": 1 }, "train": { "amp": true, "bank_dir": "data/raw/mindweave/bank-transactions-us", "batch_size": 16, "checkpoint": "runs/reconciliation-20260723/best.pt", "confidence_threshold": 0.9, "dataset": "mindweave-paired-with-synthetic-retention", "description_dropout": 0.1, "device": "cuda", "device_resolved": "cuda", "eval_every": 25, "grad_clip": 1.0, "initial_checkpoint_sha256": "e320f5cb15c7af94b803b61f20568b8e09811a0ea28bcb558984c821350ed810", "initial_checkpoint_step": 900, "learning_rate": 5e-05, "ledger_dir": "data/raw/mindweave/accounting-ledger-us", "log_every": 25, "model_parameters": 1665352, "num_workers": 0, "output_dir": "runs/mindweave-robust-20260723", "reference_dropout": 0.25, "seed": 41, "segments": 4, "source_module_dropout": 0.05, "steps": 150, "synthetic_every": 4, "synthetic_samples": 512, "train_end": "2025-06", "validation_end": "2025-09", "warmup_updates": 40, "weight_decay": 0.01 } }