Upload folder using huggingface_hub

README.md

@@ -19,90 +19,59 @@ pipeline_tag: text-classification
 
 # DistilBERT US Bank Transaction Classifier v2
 
-**Successor to [v1](https://huggingface.co/DoDataThings/distilbert-us-transaction-classifier).** Sign-aware classification with expanded merchant coverage, PayPal wrapper handling, and a refined 16-category taxonomy.
+A fine-tuned DistilBERT model that classifies US bank transaction descriptions into 17 spending categories. Uses a `[debit]`/`[credit]` sign prefix to disambiguate transaction direction — a payroll deposit and a Venmo payment look similar in text but mean opposite things financially.
 
-## What Changed from v1
+**Successor to [v1](https://huggingface.co/DoDataThings/distilbert-us-transaction-classifier)**, which classified on description text alone. v2 adds sign-aware input, expanded merchant coverage (500+), multi-format training across 8 bank statement structures, and PayPal as a first-class format.
 
-| | v1 | v2 |
-|---|---|---|
-| **Input format** | Description only | `[debit]` / `[credit]` prefix + full description |
-| **Categories** | 16 (includes Housing) | 16 (Housing split into Rent + Mortgage removed) |
-| **Training data** | 16,000 samples | 24,000 samples |
-| **Merchant coverage** | ~300 merchants | ~500+ merchants |
-| **PayPal awareness** | Limited | Full — PreApproved, Express Checkout, PP*, PAYPAL * |
-| **POS prefix awareness** | SQ* only | SQ*, TST* (Toast), CLV* (Clover) |
-| **Transfer patterns** | Basic | Brokerage sweeps, fintech platforms, wire, cashier's checks, ATM |
-
-### Why v2?
-
-v1 confused Income and Transfer because it only saw the description text. A "VENMO CASHOUT" deposit looks the same regardless of direction. v2 prepends `[credit]` or `[debit]` based on the transaction sign (after normalization to cardholder perspective), giving the model a strong directional signal.
-
-Mortgage was removed as a model category because mortgage account transactions are better classified by account type — every transaction on a mortgage account is a mortgage payment by definition.
-
-## What This Is (and Isn't)
-
-A fine-tuned DistilBERT model for classifying US bank transaction descriptions into 16 spending categories. Designed as a **fallback layer** in a multi-tier classification pipeline — not a standalone classifier.
+## How It Works
 
-1. **User rules** — pattern matching catches known merchants (highest accuracy)
-2. **This model** — classifies everything else, with sign awareness
-3. **Bank-provided categories** — fallback when model confidence is low
-4. **User overrides** — manual corrections for edge cases
+The model takes a sign prefix + transaction description and outputs one of 17 categories:
 
-## Training
-
-```
-Model:       DistilBERT-base-uncased + LoRA (r=32, alpha=64)
-Dataset:     24,000 synthetic samples, 1,500 per category
-Trainable:   1.8M / 68.7M parameters (2.6%)
-Training:    20 epochs, ~8 minutes on consumer GPU
-Best epoch:  17 (99.1% validation accuracy)
 ```
+Input:  "[debit] STARBUCKS #1234 SAN FRANCISCO CA"
+Output: Restaurants (0.99)
 
-### Loss Curve
+Input:  "[credit] Patelco CU       PAYROLL                    PPD ID: 321076470"
+Output: Income (1.00)
 
-| Epoch | Train Loss | Val Loss | Train Acc | Val Acc |
-|-------|-----------|----------|-----------|---------|
-| 1     | 2.590     | 1.936    | 20.8%     | 52.6%   |
-| 5     | 0.325     | 0.214    | 90.5%     | 93.9%   |
-| 10    | 0.078     | 0.055    | 97.8%     | 98.3%   |
-| 15    | 0.034     | 0.026    | 99.0%     | 99.2%   |
-| 17    | 0.029     | 0.030    | 99.1%     | 99.1%   |
+Input:  "[debit] CHASE CREDIT CRD AUTOPAY                    PPD ID: 4760039224"
+Output: Transfer (1.00)
 
-### Honest Assessment
-
-Validation accuracy (99.1%) is on synthetic data. Real-world performance on ~2,000 transactions:
+Input:  "[debit] PreApproved Payment Bill User Payment: Netflix"
+Output: Subscription (1.00)
+```
 
-- **86% of model classifications at 0.90+ confidence**
-- **< 0.4% below 0.50 confidence** (9 of 2,038 transactions)
-- Income and Transfer both at 100% on synthetic validation
-- Shopping remains the weakest category (~93%) due to overlap with Subscription and Groceries
+The sign prefix encodes the transaction direction from the cardholder's perspective:
+- `[debit]` — money left the account (purchases, payments out, fees)
+- `[credit]` — money entered the account (income, refunds, payments received)
 
-The sign prefix resolved the Income/Transfer confusion from v1. The main remaining challenge is niche merchants the model hasn't seen — diminishing returns territory best handled by user rules.
+This is critical for distinguishing Income from Transfer. `[credit] VENMO CASHOUT` is Income (money arriving). `[debit] VENMO PAYMENT TO JOHN SMITH` is Transfer (money leaving). The description alone can't tell you which.
 
-## Categories (16)
+## Categories (17)
 
 | Category | What it covers |
 |----------|----------------|
-| Restaurants | Fast food, sit-down, coffee shops, food delivery, POS systems (TST*, SQ*, CLV*) |
+| Restaurants | Fast food, sit-down, coffee, delivery, POS systems (TST*, SQ*, CLV*) |
 | Groceries | Supermarkets, warehouse clubs, farmers markets, convenience stores |
-| Shopping | Retail, online purchases, department stores, pet stores, liquor stores, e-commerce marketplaces |
-| Transportation | Gas, EV charging, rideshare, auto maintenance, parking, tolls, DMV |
-| Entertainment | Movies, events, gaming (Steam, PlayStation), gambling/sportsbooks |
+| Shopping | Retail, online, department stores, pet stores, liquor stores, e-commerce marketplaces |
+| Transportation | Gas, EV charging, rideshare, auto service, parking, tolls, DMV |
+| Entertainment | Movies, events, gaming, gambling/sportsbooks |
 | Utilities | Electric, internet, phone, water, waste/trash, solar |
-| Subscription | Streaming, SaaS, AI tools, VPNs, social media premium, dating apps, news |
+| Subscription | Streaming, SaaS, AI tools, VPNs, social media premium, dating, business SaaS |
 | Healthcare | Pharmacy, doctor, dentist, telehealth, vision, hospital |
-| Insurance | Auto, home, health, life insurance |
+| Insurance | Auto, home, health, life, home warranty |
+| Mortgage | Bank, credit union, and fintech mortgage payments, escrow, principal |
 | Rent | Property management companies, lease payments |
 | Travel | Hotels, airlines, car rental, cruise lines, airport services |
 | Education | Online courses, tutoring, books, tuition, certification |
 | Personal Care | Salon, gym, beauty, spa, barber |
-| Transfer | CC autopay, Zelle/Venmo sends, bank transfers, brokerage sweeps, BNPL, wire transfers, ATM, cashier's checks |
+| Transfer | CC autopay, P2P sends, bank transfers, brokerage sweeps, fintech, BNPL, wire, ATM, cashier's checks |
 | Income | Payroll, direct deposit, interest, refunds, government benefits, gig economy payouts |
-| Fees | Bank fees, late fees, service charges, ATM fees |
+| Fees | Bank fees, late fees, ATM surcharges, service charges |
 
-### Account-Type-Implied Categories (not model-classified)
+### Account-Type-Implied Categories
 
-These categories are determined by the account type, not the model:
+If you know the account type, some categories can be assigned without the model:
 
 | Account Type | Category |
 |---|---|
@@ -113,6 +82,45 @@ These categories are determined by the account type, not the model:
 | HELOC | Transfer |
 | CD | Income |
 
+For checking, savings, and credit card accounts, use the model.
+
+## Training
+
+```
+Model:       DistilBERT-base-uncased + LoRA (r=32, alpha=64)
+Dataset:     68,000 synthetic samples (4,000 per category)
+Trainable:   1.8M / 68.7M parameters (2.6%)
+Training:    20 epochs, best at epoch 16
+Validation:  99.9% accuracy (15 of 17 categories at 100%)
+```
+
+### Multi-Format Training
+
+The model is trained on 8 bank statement formats so it classifies correctly regardless of which bank produced the description:
+
+| Format | Example | Source |
+|---|---|---|
+| Chase merchant | `STARBUCKS #1234` | Chase credit cards |
+| Chase ACH | `INSTITUTION     PURPOSE        PPD ID: CODE` | Chase checking |
+| Apple Card | `MERCHANT ADDRESS CITY ZIP STATE USA` | Apple Card |
+| PayPal native | `PreApproved Payment Bill User Payment: MERCHANT` | PayPal credit card |
+| PayPal prefix | `PP*MERCHANT`, `PYPL*MERCHANT`, `PAYPAL *MERCHANT` | Chase/other banks |
+| Capital One | `Withdrawal from MERCHANT`, `Preauthorized Deposit from MERCHANT` | Capital One |
+| Mercury | `MERCHANT; Description` or just `MERCHANT` | Mercury, neobanks |
+| POS prefix | `SQ *MERCHANT`, `TST*MERCHANT`, `CLV*MERCHANT` | Square, Toast, Clover |
+
+PayPal formats appear across all spending categories at meaningful rates, reflecting that people use PayPal cards at any merchant.
+
+### Honest Assessment
+
+The 99.9% validation accuracy is on synthetic data. On ~2,000 real transactions:
+
+- **96.1% of model classifications at 0.90+ confidence**
+- **< 0.5% below 0.50 confidence**
+- 17 bank-category fallbacks (obscure merchants where the model defers)
+- Shopping is the weakest category due to overlap with Subscription and Groceries
+- Niche/unknown merchants may classify with lower confidence — use merchant rules for known edge cases
+
 ## Usage
 
 ### Python
@@ -122,9 +130,9 @@ from transformers import pipeline
 
 classifier = pipeline("text-classification", model="DoDataThings/distilbert-us-transaction-classifier-v2")
 
-# v2 requires sign prefix
+# Sign prefix required
 result = classifier("[debit] STARBUCKS #1234 SAN FRANCISCO CA")
-print(result)  # [{'label': 'Restaurants', 'score': 0.98}]
+print(result)  # [{'label': 'Restaurants', 'score': 0.99}]
 
 # Sign matters for ambiguous transactions
 classifier("[credit] VENMO CASHOUT PPD ID: 12345678")
@@ -132,6 +140,13 @@ classifier("[credit] VENMO CASHOUT PPD ID: 12345678")
 
 classifier("[debit] VENMO PAYMENT TO JOHN SMITH")
 # [{'label': 'Transfer', 'score': 0.97}]
+
+# Works across all bank formats
+classifier("[debit] PreApproved Payment Bill User Payment: Netflix")
+# [{'label': 'Subscription', 'score': 1.00}]
+
+classifier("[debit] PP*SAFEWAY")
+# [{'label': 'Groceries', 'score': 1.00}]
 ```
 
 ### JavaScript (Transformers.js)
@@ -145,31 +160,39 @@ const classifier = await pipeline(
 );
 
 const result = await classifier('[debit] STARBUCKS #1234');
-// [{ label: 'Restaurants', score: 0.98 }]
+// [{ label: 'Restaurants', score: 0.99 }]
 ```
 
 An ONNX export is included in the `onnx/` subdirectory.
 
-### Sign Prefix Convention
-
-Prepend `[credit]` or `[debit]` based on the **normalized** transaction amount (cardholder perspective):
-- `[debit]` — money left the account (purchases, payments out, fees)
-- `[credit]` — money entered the account (income, refunds, payments received)
+## Design Decisions
 
-If your data uses issuer perspective (e.g., Apple Card where purchases are positive), normalize the sign first, then apply the prefix.
+- **Sign prefix, not account type.** We considered passing account type (checking, credit, etc.) as a feature but concluded that sign alone provides the disambiguation signal. Account type is an upstream routing concern — it determines which classifier runs, not what the classifier outputs.
+- **17 model categories + 6 account-type categories.** Mortgage is both a model category (for classifying mortgage descriptions on checking accounts) and an account-type-implied category (for mortgage account transactions). This serves both use cases — people with account type metadata and people with just transaction descriptions.
+- **PayPal as a bank format, not a wrapper.** PayPal is a card issuer. People use PayPal cards at restaurants, grocery stores, and everywhere else. The training data treats PayPal formats as first-class bank statement structures across all categories.
+- **Synthetic data with real formats.** The training data is synthetic but models real bank statement patterns — Chase ACH padding, Apple Card address formats, Capital One action prefixes, Mercury's minimal format. The generator is open source so you can extend it.
 
 ## Training Data
 
-The synthetic dataset is published at [`DoDataThings/us-bank-transaction-categories-v2`](https://huggingface.co/datasets/DoDataThings/us-bank-transaction-categories-v2). The generator script is open source — you can extend the merchant pools, add format templates, or increase sample counts.
+The dataset is published at [`DoDataThings/us-bank-transaction-categories-v2`](https://huggingface.co/datasets/DoDataThings/us-bank-transaction-categories-v2).
+
+## Generator
+
+The synthetic data generator is open source:
+
+```bash
+node scripts/generate-training-data.js --count 4000  # 4,000 per category
+```
+
+Available at [github.com/wnstnb/foliome](https://github.com/wnstnb/foliome).
 
 ## Limitations
 
-- **US bank formats only** — Trained on Chase, Apple Card, PayPal, Capital One, and US Bank statement patterns
+- **US bank formats only** — Trained on Chase, Apple Card, PayPal, Capital One, Mercury, and US Bank patterns
 - **Synthetic training data** — May miss patterns from banks not represented
-- **Shopping is the weakest category** (~93%) due to overlap with Subscription and Groceries
-- **Niche merchants** may classify with low confidence — use merchant rules for known edge cases
-- **Sign prefix required** — The model expects `[debit]` or `[credit]` prefix. Passing raw descriptions without the prefix will degrade accuracy.
-- **Not a standalone solution** — Best results come from combining with merchant rules and account-type-implied classifications
+- **Shopping is the weakest category** due to overlap with Subscription and Groceries
+- **Sign prefix required** — Passing raw descriptions without `[debit]`/`[credit]` will degrade accuracy
+- **Not a standalone solution** — Best results come from combining with merchant rules and account-type classification
 
 ## License
 

config.json

@@ -16,15 +16,16 @@
     "4": "Healthcare",
     "5": "Income",
     "6": "Insurance",
-    "7": "Personal Care",
-    "8": "Rent",
-    "9": "Restaurants",
-    "10": "Shopping",
-    "11": "Subscription",
-    "12": "Transfer",
-    "13": "Transportation",
-    "14": "Travel",
-    "15": "Utilities"
+    "7": "Mortgage",
+    "8": "Personal Care",
+    "9": "Rent",
+    "10": "Restaurants",
+    "11": "Shopping",
+    "12": "Subscription",
+    "13": "Transfer",
+    "14": "Transportation",
+    "15": "Travel",
+    "16": "Utilities"
   },
   "initializer_range": 0.02,
   "label2id": {
@@ -35,15 +36,16 @@
     "Healthcare": 4,
     "Income": 5,
     "Insurance": 6,
-    "Personal Care": 7,
-    "Rent": 8,
-    "Restaurants": 9,
-    "Shopping": 10,
-    "Subscription": 11,
-    "Transfer": 12,
-    "Transportation": 13,
-    "Travel": 14,
-    "Utilities": 15
+    "Mortgage": 7,
+    "Personal Care": 8,
+    "Rent": 9,
+    "Restaurants": 10,
+    "Shopping": 11,
+    "Subscription": 12,
+    "Transfer": 13,
+    "Transportation": 14,
+    "Travel": 15,
+    "Utilities": 16
   },
   "max_position_embeddings": 512,
   "model_type": "distilbert",

label_mapping.json

@@ -7,15 +7,16 @@
     "4": "Healthcare",
     "5": "Income",
     "6": "Insurance",
-    "7": "Personal Care",
-    "8": "Rent",
-    "9": "Restaurants",
-    "10": "Shopping",
-    "11": "Subscription",
-    "12": "Transfer",
-    "13": "Transportation",
-    "14": "Travel",
-    "15": "Utilities"
+    "7": "Mortgage",
+    "8": "Personal Care",
+    "9": "Rent",
+    "10": "Restaurants",
+    "11": "Shopping",
+    "12": "Subscription",
+    "13": "Transfer",
+    "14": "Transportation",
+    "15": "Travel",
+    "16": "Utilities"
   },
   "label2id": {
     "Education": 0,
@@ -25,14 +26,15 @@
     "Healthcare": 4,
     "Income": 5,
     "Insurance": 6,
-    "Personal Care": 7,
-    "Rent": 8,
-    "Restaurants": 9,
-    "Shopping": 10,
-    "Subscription": 11,
-    "Transfer": 12,
-    "Transportation": 13,
-    "Travel": 14,
-    "Utilities": 15
+    "Mortgage": 7,
+    "Personal Care": 8,
+    "Rent": 9,
+    "Restaurants": 10,
+    "Shopping": 11,
+    "Subscription": 12,
+    "Transfer": 13,
+    "Transportation": 14,
+    "Travel": 15,
+    "Utilities": 16
   }
 }

onnx/model_quantized.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:af4d35409501558e9112dbc5aef014f0c8086427d00b4950d629f497d20d54fd
-size 267975237
+oid sha256:1ab2b64277921566be0303003473fac0bb6b65b21dd7db74548575410eecc5c6
+size 267978313

training-report.json

@@ -1,14 +1,14 @@
 {
   "model": "distilbert-base-uncased + LoRA (r=32, alpha=64)",
-  "dataset": "24000 synthetic transactions, 16 categories, 1500/category",
-  "split": "20400 train / 3600 val (85/15)",
+  "dataset": "68000 synthetic transactions, 17 categories, 4000/category",
+  "split": "57800 train / 10200 val (85/15)",
   "epochs": 20,
-  "best_epoch": 18,
-  "best_val_accuracy": 0.9908,
-  "total_training_time_s": 458.7,
+  "best_epoch": 16,
+  "best_val_accuracy": 0.9992,
+  "total_training_time_s": 1262.2,
   "device": "cuda",
-  "trainable_params": 1782544,
-  "total_params": 68748320,
+  "trainable_params": 1783313,
+  "total_params": 68749858,
   "categories": [
     "Education",
     "Entertainment",
@@ -17,6 +17,7 @@
     "Healthcare",
     "Income",
     "Insurance",
+    "Mortgage",
     "Personal Care",
     "Rent",
     "Restaurants",
@@ -30,643 +31,651 @@
   "history": [
     {
       "epoch": 1,
-      "train_loss": 2.5896,
-      "val_loss": 1.9356,
-      "train_acc": 0.2075,
-      "val_acc": 0.5256,
+      "train_loss": 2.197,
+      "val_loss": 1.1123,
+      "train_acc": 0.3841,
+      "val_acc": 0.713,
       "per_category": {
-        "Education": 0.362,
-        "Entertainment": 0.142,
-        "Fees": 1.0,
-        "Groceries": 0.575,
-        "Healthcare": 0.798,
-        "Income": 0.964,
-        "Insurance": 0.273,
-        "Personal Care": 0.203,
-        "Rent": 0.97,
-        "Restaurants": 0.544,
-        "Shopping": 0.36,
-        "Subscription": 0.802,
-        "Transfer": 0.166,
-        "Transportation": 0.5,
-        "Travel": 0.241,
-        "Utilities": 0.479
+        "Education": 0.649,
+        "Entertainment": 0.644,
+        "Fees": 0.995,
+        "Groceries": 0.419,
+        "Healthcare": 0.815,
+        "Income": 0.979,
+        "Insurance": 0.66,
+        "Mortgage": 0.901,
+        "Personal Care": 0.535,
+        "Rent": 0.98,
+        "Restaurants": 0.664,
+        "Shopping": 0.522,
+        "Subscription": 0.749,
+        "Transfer": 0.653,
+        "Transportation": 0.595,
+        "Travel": 0.571,
+        "Utilities": 0.776
       },
-      "epoch_time_s": 23.7
+      "epoch_time_s": 62.7
     },
     {
       "epoch": 2,
-      "train_loss": 1.3952,
-      "val_loss": 0.8527,
-      "train_acc": 0.6386,
-      "val_acc": 0.7642,
+      "train_loss": 0.7792,
+      "val_loss": 0.3999,
+      "train_acc": 0.7801,
+      "val_acc": 0.879,
       "per_category": {
-        "Education": 0.819,
-        "Entertainment": 0.652,
-        "Fees": 1.0,
-        "Groceries": 0.617,
-        "Healthcare": 0.88,
-        "Income": 1.0,
-        "Insurance": 0.745,
-        "Personal Care": 0.586,
-        "Rent": 0.991,
-        "Restaurants": 0.628,
-        "Shopping": 0.671,
-        "Subscription": 0.718,
-        "Transfer": 0.751,
-        "Transportation": 0.71,
-        "Travel": 0.565,
-        "Utilities": 0.853
+        "Education": 0.966,
+        "Entertainment": 0.876,
+        "Fees": 0.997,
+        "Groceries": 0.7,
+        "Healthcare": 0.922,
+        "Income": 0.997,
+        "Insurance": 0.906,
+        "Mortgage": 0.982,
+        "Personal Care": 0.986,
+        "Rent": 1.0,
+        "Restaurants": 0.841,
+        "Shopping": 0.642,
+        "Subscription": 0.792,
+        "Transfer": 0.873,
+        "Transportation": 0.803,
+        "Travel": 0.765,
+        "Utilities": 0.899
       },
-      "epoch_time_s": 23.2
+      "epoch_time_s": 62.9
     },
     {
       "epoch": 3,
-      "train_loss": 0.7496,
-      "val_loss": 0.4969,
-      "train_acc": 0.7876,
-      "val_acc": 0.8619,
+      "train_loss": 0.3276,
+      "val_loss": 0.134,
+      "train_acc": 0.9017,
+      "val_acc": 0.963,
       "per_category": {
-        "Education": 0.95,
-        "Entertainment": 0.826,
+        "Education": 0.988,
+        "Entertainment": 0.978,
         "Fees": 1.0,
-        "Groceries": 0.696,
-        "Healthcare": 0.93,
-        "Income": 0.982,
-        "Insurance": 0.892,
-        "Personal Care": 0.918,
+        "Groceries": 0.915,
+        "Healthcare": 0.981,
+        "Income": 0.995,
+        "Insurance": 0.997,
+        "Mortgage": 0.989,
+        "Personal Care": 0.997,
         "Rent": 1.0,
-        "Restaurants": 0.686,
-        "Shopping": 0.778,
-        "Subscription": 0.861,
-        "Transfer": 0.889,
-        "Transportation": 0.694,
-        "Travel": 0.728,
-        "Utilities": 0.943
+        "Restaurants": 0.917,
+        "Shopping": 0.841,
+        "Subscription": 0.893,
+        "Transfer": 0.988,
+        "Transportation": 0.922,
+        "Travel": 0.983,
+        "Utilities": 0.987
       },
-      "epoch_time_s": 23.5
+      "epoch_time_s": 62.8
     },
     {
       "epoch": 4,
-      "train_loss": 0.4773,
-      "val_loss": 0.2947,
-      "train_acc": 0.8612,
-      "val_acc": 0.9169,
+      "train_loss": 0.1417,
+      "val_loss": 0.0588,
+      "train_acc": 0.9577,
+      "val_acc": 0.9824,
       "per_category": {
-        "Education": 0.977,
-        "Entertainment": 0.939,
+        "Education": 0.997,
+        "Entertainment": 0.998,
         "Fees": 1.0,
-        "Groceries": 0.734,
-        "Healthcare": 0.946,
-        "Income": 1.0,
-        "Insurance": 0.944,
-        "Personal Care": 0.991,
+        "Groceries": 1.0,
+        "Healthcare": 0.982,
+        "Income": 0.997,
+        "Insurance": 0.994,
+        "Mortgage": 0.998,
+        "Personal Care": 1.0,
         "Rent": 1.0,
-        "Restaurants": 0.863,
-        "Shopping": 0.796,
-        "Subscription": 0.842,
-        "Transfer": 0.926,
-        "Transportation": 0.863,
-        "Travel": 0.869,
-        "Utilities": 0.953
+        "Restaurants": 0.98,
+        "Shopping": 0.871,
+        "Subscription": 0.918,
+        "Transfer": 0.98,
+        "Transportation": 0.991,
+        "Travel": 0.997,
+        "Utilities": 0.998
       },
-      "epoch_time_s": 23.4
+      "epoch_time_s": 63.0
     },
     {
       "epoch": 5,
-      "train_loss": 0.325,
-      "val_loss": 0.2142,
-      "train_acc": 0.9049,
-      "val_acc": 0.9392,
+      "train_loss": 0.0701,
+      "val_loss": 0.0272,
+      "train_acc": 0.9788,
+      "val_acc": 0.9925,
       "per_category": {
-        "Education": 0.982,
-        "Entertainment": 0.951,
+        "Education": 1.0,
+        "Entertainment": 0.995,
         "Fees": 1.0,
-        "Groceries": 0.883,
-        "Healthcare": 0.967,
+        "Groceries": 1.0,
+        "Healthcare": 0.998,
         "Income": 1.0,
-        "Insurance": 0.97,
+        "Insurance": 0.998,
+        "Mortgage": 0.996,
         "Personal Care": 1.0,
         "Rent": 1.0,
-        "Restaurants": 0.881,
-        "Shopping": 0.804,
-        "Subscription": 0.866,
-        "Transfer": 0.917,
-        "Transportation": 0.891,
-        "Travel": 0.921,
-        "Utilities": 0.976
+        "Restaurants": 0.995,
+        "Shopping": 0.919,
+        "Subscription": 0.98,
+        "Transfer": 0.995,
+        "Transportation": 0.995,
+        "Travel": 1.0,
+        "Utilities": 1.0
       },
-      "epoch_time_s": 23.4
+      "epoch_time_s": 63.3
     },
     {
       "epoch": 6,
-      "train_loss": 0.2342,
-      "val_loss": 0.1447,
-      "train_acc": 0.9325,
-      "val_acc": 0.9575,
+      "train_loss": 0.042,
+      "val_loss": 0.0126,
+      "train_acc": 0.9873,
+      "val_acc": 0.9967,
       "per_category": {
-        "Education": 0.982,
-        "Entertainment": 0.988,
+        "Education": 1.0,
+        "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 0.883,
-        "Healthcare": 0.967,
+        "Groceries": 1.0,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.939,
+        "Insurance": 1.0,
+        "Mortgage": 0.998,
         "Personal Care": 1.0,
         "Rent": 1.0,
-        "Restaurants": 0.947,
-        "Shopping": 0.836,
-        "Subscription": 0.941,
-        "Transfer": 0.982,
-        "Transportation": 0.907,
-        "Travel": 0.948,
-        "Utilities": 0.995
+        "Restaurants": 1.0,
+        "Shopping": 0.968,
+        "Subscription": 0.985,
+        "Transfer": 0.997,
+        "Transportation": 0.995,
+        "Travel": 1.0,
+        "Utilities": 1.0
       },
-      "epoch_time_s": 23.4
+      "epoch_time_s": 63.3
     },
     {
       "epoch": 7,
-      "train_loss": 0.1724,
-      "val_loss": 0.1079,
-      "train_acc": 0.9508,
-      "val_acc": 0.9708,
+      "train_loss": 0.0254,
+      "val_loss": 0.0114,
+      "train_acc": 0.9924,
+      "val_acc": 0.9973,
       "per_category": {
-        "Education": 0.982,
-        "Entertainment": 0.996,
+        "Education": 1.0,
+        "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 0.953,
-        "Healthcare": 0.967,
+        "Groceries": 1.0,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.991,
-        "Personal Care": 1.0,
+        "Insurance": 1.0,
+        "Mortgage": 0.998,
+        "Personal Care": 0.998,
         "Rent": 1.0,
-        "Restaurants": 0.942,
-        "Shopping": 0.844,
-        "Subscription": 0.926,
-        "Transfer": 0.982,
-        "Transportation": 0.972,
-        "Travel": 0.979,
-        "Utilities": 0.991
+        "Restaurants": 0.992,
+        "Shopping": 0.972,
+        "Subscription": 0.998,
+        "Transfer": 1.0,
+        "Transportation": 0.995,
+        "Travel": 1.0,
+        "Utilities": 1.0
       },
-      "epoch_time_s": 23.5
+      "epoch_time_s": 63.2
     },
     {
       "epoch": 8,
-      "train_loss": 0.1254,
-      "val_loss": 0.0775,
-      "train_acc": 0.9649,
-      "val_acc": 0.9789,
+      "train_loss": 0.0202,
+      "val_loss": 0.0072,
+      "train_acc": 0.994,
+      "val_acc": 0.9983,
       "per_category": {
         "Education": 1.0,
-        "Entertainment": 0.988,
+        "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 0.967,
-        "Healthcare": 0.967,
+        "Groceries": 1.0,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.996,
+        "Insurance": 1.0,
+        "Mortgage": 0.998,
         "Personal Care": 1.0,
         "Rent": 1.0,
-        "Restaurants": 0.96,
-        "Shopping": 0.893,
-        "Subscription": 0.95,
-        "Transfer": 0.982,
-        "Transportation": 0.976,
-        "Travel": 0.995,
-        "Utilities": 0.986
+        "Restaurants": 1.0,
+        "Shopping": 0.978,
+        "Subscription": 0.998,
+        "Transfer": 1.0,
+        "Transportation": 0.997,
+        "Travel": 1.0,
+        "Utilities": 1.0
       },
-      "epoch_time_s": 23.5
+      "epoch_time_s": 63.2
     },
     {
       "epoch": 9,
-      "train_loss": 0.0972,
-      "val_loss": 0.0627,
-      "train_acc": 0.9723,
-      "val_acc": 0.9814,
+      "train_loss": 0.0139,
+      "val_loss": 0.0035,
+      "train_acc": 0.9958,
+      "val_acc": 0.999,
       "per_category": {
-        "Education": 0.995,
-        "Entertainment": 0.992,
+        "Education": 1.0,
+        "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 0.981,
-        "Healthcare": 0.967,
+        "Groceries": 1.0,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.996,
+        "Insurance": 1.0,
+        "Mortgage": 1.0,
         "Personal Care": 1.0,
         "Rent": 1.0,
-        "Restaurants": 0.96,
-        "Shopping": 0.884,
-        "Subscription": 0.96,
-        "Transfer": 0.982,
-        "Transportation": 0.988,
-        "Travel": 0.995,
+        "Restaurants": 1.0,
+        "Shopping": 0.983,
+        "Subscription": 1.0,
+        "Transfer": 1.0,
+        "Transportation": 1.0,
+        "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.4
+      "epoch_time_s": 63.3
     },
     {
       "epoch": 10,
-      "train_loss": 0.0771,
-      "val_loss": 0.0557,
-      "train_acc": 0.9789,
-      "val_acc": 0.985,
+      "train_loss": 0.0121,
+      "val_loss": 0.0043,
+      "train_acc": 0.9966,
+      "val_acc": 0.9989,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
         "Groceries": 1.0,
-        "Healthcare": 0.963,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.996,
-        "Personal Care": 1.0,
+        "Insurance": 1.0,
+        "Mortgage": 1.0,
+        "Personal Care": 0.998,
         "Rent": 1.0,
-        "Restaurants": 0.978,
-        "Shopping": 0.889,
-        "Subscription": 0.955,
-        "Transfer": 0.995,
-        "Transportation": 0.988,
-        "Travel": 0.995,
+        "Restaurants": 1.0,
+        "Shopping": 0.987,
+        "Subscription": 0.998,
+        "Transfer": 1.0,
+        "Transportation": 0.998,
+        "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.6
+      "epoch_time_s": 63.4
     },
     {
       "epoch": 11,
-      "train_loss": 0.0665,
-      "val_loss": 0.0485,
-      "train_acc": 0.9812,
-      "val_acc": 0.9864,
+      "train_loss": 0.0094,
+      "val_loss": 0.0049,
+      "train_acc": 0.9975,
+      "val_acc": 0.9988,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 0.991,
-        "Healthcare": 0.967,
+        "Groceries": 1.0,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.996,
+        "Insurance": 1.0,
+        "Mortgage": 1.0,
         "Personal Care": 1.0,
         "Rent": 1.0,
-        "Restaurants": 0.987,
-        "Shopping": 0.898,
-        "Subscription": 0.975,
-        "Transfer": 0.982,
-        "Transportation": 0.988,
+        "Restaurants": 1.0,
+        "Shopping": 0.982,
+        "Subscription": 1.0,
+        "Transfer": 1.0,
+        "Transportation": 0.998,
         "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.6
+      "epoch_time_s": 63.2
     },
     {
       "epoch": 12,
-      "train_loss": 0.0534,
-      "val_loss": 0.0404,
-      "train_acc": 0.9841,
-      "val_acc": 0.9872,
+      "train_loss": 0.0087,
+      "val_loss": 0.0027,
+      "train_acc": 0.9976,
+      "val_acc": 0.9991,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 0.986,
-        "Healthcare": 0.963,
+        "Groceries": 1.0,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.996,
-        "Personal Care": 0.991,
+        "Insurance": 1.0,
+        "Mortgage": 1.0,
+        "Personal Care": 0.998,
         "Rent": 1.0,
-        "Restaurants": 0.991,
-        "Shopping": 0.907,
-        "Subscription": 0.98,
-        "Transfer": 0.995,
-        "Transportation": 0.988,
+        "Restaurants": 1.0,
+        "Shopping": 0.988,
+        "Subscription": 1.0,
+        "Transfer": 1.0,
+        "Transportation": 0.998,
         "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.6
+      "epoch_time_s": 63.3
     },
     {
       "epoch": 13,
-      "train_loss": 0.0463,
-      "val_loss": 0.0418,
-      "train_acc": 0.9857,
-      "val_acc": 0.9889,
+      "train_loss": 0.0068,
+      "val_loss": 0.0047,
+      "train_acc": 0.9981,
+      "val_acc": 0.9988,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 1.0,
-        "Healthcare": 0.967,
+        "Groceries": 0.998,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.996,
+        "Insurance": 1.0,
+        "Mortgage": 1.0,
         "Personal Care": 1.0,
         "Rent": 1.0,
-        "Restaurants": 0.991,
-        "Shopping": 0.898,
-        "Subscription": 0.975,
-        "Transfer": 0.995,
-        "Transportation": 1.0,
+        "Restaurants": 1.0,
+        "Shopping": 0.983,
+        "Subscription": 1.0,
+        "Transfer": 1.0,
+        "Transportation": 0.998,
         "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.3
+      "epoch_time_s": 63.1
     },
     {
       "epoch": 14,
-      "train_loss": 0.0421,
-      "val_loss": 0.0386,
-      "train_acc": 0.9872,
-      "val_acc": 0.9889,
+      "train_loss": 0.0066,
+      "val_loss": 0.0029,
+      "train_acc": 0.9981,
+      "val_acc": 0.9989,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 0.995,
-        "Healthcare": 0.967,
+        "Groceries": 0.998,
+        "Healthcare": 1.0,
         "Income": 1.0,
         "Insurance": 1.0,
-        "Personal Care": 1.0,
+        "Mortgage": 1.0,
+        "Personal Care": 0.998,
         "Rent": 1.0,
-        "Restaurants": 0.991,
-        "Shopping": 0.911,
-        "Subscription": 0.98,
-        "Transfer": 0.991,
-        "Transportation": 0.988,
+        "Restaurants": 1.0,
+        "Shopping": 0.988,
+        "Subscription": 1.0,
+        "Transfer": 1.0,
+        "Transportation": 0.997,
         "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.3
+      "epoch_time_s": 63.1
     },
     {
       "epoch": 15,
-      "train_loss": 0.0378,
-      "val_loss": 0.0341,
-      "train_acc": 0.9886,
-      "val_acc": 0.9892,
+      "train_loss": 0.0059,
+      "val_loss": 0.0049,
+      "train_acc": 0.9983,
+      "val_acc": 0.9989,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 0.995,
-        "Healthcare": 0.971,
+        "Groceries": 1.0,
+        "Healthcare": 1.0,
         "Income": 1.0,
         "Insurance": 1.0,
+        "Mortgage": 1.0,
         "Personal Care": 1.0,
         "Rent": 1.0,
-        "Restaurants": 0.991,
-        "Shopping": 0.907,
-        "Subscription": 0.98,
-        "Transfer": 0.995,
-        "Transportation": 0.988,
+        "Restaurants": 1.0,
+        "Shopping": 0.985,
+        "Subscription": 1.0,
+        "Transfer": 1.0,
+        "Transportation": 0.997,
         "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.4
+      "epoch_time_s": 63.0
     },
     {
       "epoch": 16,
-      "train_loss": 0.0319,
-      "val_loss": 0.0363,
-      "train_acc": 0.9912,
-      "val_acc": 0.9894,
+      "train_loss": 0.0055,
+      "val_loss": 0.0033,
+      "train_acc": 0.9983,
+      "val_acc": 0.9992,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
         "Groceries": 1.0,
-        "Healthcare": 0.967,
+        "Healthcare": 1.0,
         "Income": 1.0,
         "Insurance": 1.0,
-        "Personal Care": 1.0,
+        "Mortgage": 1.0,
+        "Personal Care": 0.998,
         "Rent": 1.0,
-        "Restaurants": 0.991,
-        "Shopping": 0.902,
-        "Subscription": 0.985,
-        "Transfer": 0.995,
-        "Transportation": 0.992,
+        "Restaurants": 1.0,
+        "Shopping": 0.988,
+        "Subscription": 1.0,
+        "Transfer": 1.0,
+        "Transportation": 1.0,
         "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.4
+      "epoch_time_s": 63.1
     },
     {
       "epoch": 17,
-      "train_loss": 0.0288,
-      "val_loss": 0.0296,
-      "train_acc": 0.9913,
-      "val_acc": 0.9906,
+      "train_loss": 0.0049,
+      "val_loss": 0.0032,
+      "train_acc": 0.9987,
+      "val_acc": 0.9991,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 0.986,
-        "Healthcare": 0.971,
+        "Groceries": 1.0,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.996,
-        "Personal Care": 0.991,
+        "Insurance": 1.0,
+        "Mortgage": 1.0,
+        "Personal Care": 0.998,
         "Rent": 1.0,
-        "Restaurants": 0.991,
-        "Shopping": 0.947,
-        "Subscription": 0.985,
-        "Transfer": 0.995,
-        "Transportation": 0.988,
+        "Restaurants": 1.0,
+        "Shopping": 0.987,
+        "Subscription": 1.0,
+        "Transfer": 1.0,
+        "Transportation": 1.0,
         "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.3
+      "epoch_time_s": 63.0
     },
     {
       "epoch": 18,
-      "train_loss": 0.0255,
-      "val_loss": 0.0284,
-      "train_acc": 0.993,
-      "val_acc": 0.9908,
+      "train_loss": 0.0042,
+      "val_loss": 0.0034,
+      "train_acc": 0.9988,
+      "val_acc": 0.9991,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
-        "Groceries": 0.995,
-        "Healthcare": 0.971,
+        "Groceries": 0.998,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.996,
-        "Personal Care": 1.0,
+        "Insurance": 1.0,
+        "Mortgage": 1.0,
+        "Personal Care": 0.998,
         "Rent": 1.0,
-        "Restaurants": 0.991,
-        "Shopping": 0.929,
-        "Subscription": 0.985,
-        "Transfer": 0.995,
-        "Transportation": 0.992,
+        "Restaurants": 1.0,
+        "Shopping": 0.988,
+        "Subscription": 1.0,
+        "Transfer": 1.0,
+        "Transportation": 1.0,
         "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.4
+      "epoch_time_s": 63.0
     },
     {
       "epoch": 19,
-      "train_loss": 0.0273,
-      "val_loss": 0.0306,
-      "train_acc": 0.9912,
-      "val_acc": 0.9897,
+      "train_loss": 0.0045,
+      "val_loss": 0.0032,
+      "train_acc": 0.9989,
+      "val_acc": 0.9992,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
         "Groceries": 1.0,
-        "Healthcare": 0.967,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.996,
-        "Personal Care": 1.0,
+        "Insurance": 1.0,
+        "Mortgage": 1.0,
+        "Personal Care": 0.998,
         "Rent": 1.0,
-        "Restaurants": 0.991,
-        "Shopping": 0.907,
-        "Subscription": 0.98,
+        "Restaurants": 1.0,
+        "Shopping": 0.988,
+        "Subscription": 1.0,
         "Transfer": 1.0,
-        "Transportation": 0.996,
+        "Transportation": 1.0,
         "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 22.5
+      "epoch_time_s": 63.0
     },
     {
       "epoch": 20,
-      "train_loss": 0.023,
-      "val_loss": 0.03,
-      "train_acc": 0.9928,
-      "val_acc": 0.9906,
+      "train_loss": 0.0038,
+      "val_loss": 0.0032,
+      "train_acc": 0.9989,
+      "val_acc": 0.9992,
       "per_category": {
         "Education": 1.0,
         "Entertainment": 1.0,
         "Fees": 1.0,
         "Groceries": 1.0,
-        "Healthcare": 0.967,
+        "Healthcare": 1.0,
         "Income": 1.0,
-        "Insurance": 0.996,
-        "Personal Care": 1.0,
+        "Insurance": 1.0,
+        "Mortgage": 1.0,
+        "Personal Care": 0.998,
         "Rent": 1.0,
-        "Restaurants": 0.991,
-        "Shopping": 0.916,
-        "Subscription": 0.985,
+        "Restaurants": 1.0,
+        "Shopping": 0.988,
+        "Subscription": 1.0,
         "Transfer": 1.0,
-        "Transportation": 0.996,
+        "Transportation": 1.0,
         "Travel": 1.0,
         "Utilities": 1.0
       },
-      "epoch_time_s": 23.2
+      "epoch_time_s": 63.0
     }
   ],
   "final_per_category": {
     "Education": {
       "accuracy": 1.0,
-      "correct": 221,
-      "total": 221,
+      "correct": 582,
+      "total": 582,
       "top_confusions": {}
     },
     "Entertainment": {
       "accuracy": 1.0,
-      "correct": 247,
-      "total": 247,
+      "correct": 579,
+      "total": 579,
       "top_confusions": {}
     },
     "Fees": {
       "accuracy": 1.0,
-      "correct": 240,
-      "total": 240,
+      "correct": 609,
+      "total": 609,
       "top_confusions": {}
     },
     "Groceries": {
-      "accuracy": 0.995,
-      "correct": 213,
-      "total": 214,
-      "top_confusions": {
-        "Shopping": 1
-      }
+      "accuracy": 1.0,
+      "correct": 613,
+      "total": 613,
+      "top_confusions": {}
     },
     "Healthcare": {
-      "accuracy": 0.971,
-      "correct": 235,
-      "total": 242,
-      "top_confusions": {
-        "Education": 2,
-        "Utilities": 2,
-        "Insurance": 1
-      }
+      "accuracy": 1.0,
+      "correct": 616,
+      "total": 616,
+      "top_confusions": {}
     },
     "Income": {
       "accuracy": 1.0,
-      "correct": 221,
-      "total": 221,
+      "correct": 626,
+      "total": 626,
       "top_confusions": {}
     },
     "Insurance": {
-      "accuracy": 0.996,
-      "correct": 230,
-      "total": 231,
-      "top_confusions": {
-        "Income": 1
-      }
+      "accuracy": 1.0,
+      "correct": 635,
+      "total": 635,
+      "top_confusions": {}
     },
-    "Personal Care": {
+    "Mortgage": {
       "accuracy": 1.0,
-      "correct": 232,
-      "total": 232,
+      "correct": 565,
+      "total": 565,
       "top_confusions": {}
     },
+    "Personal Care": {
+      "accuracy": 0.998,
+      "correct": 584,
+      "total": 585,
+      "top_confusions": {
+        "Shopping": 1
+      }
+    },
     "Rent": {
       "accuracy": 1.0,
-      "correct": 232,
-      "total": 232,
+      "correct": 608,
+      "total": 608,
       "top_confusions": {}
     },
     "Restaurants": {
-      "accuracy": 0.991,
-      "correct": 224,
-      "total": 226,
-      "top_confusions": {
-        "Groceries": 2
-      }
+      "accuracy": 1.0,
+      "correct": 590,
+      "total": 590,
+      "top_confusions": {}
     },
     "Shopping": {
-      "accuracy": 0.929,
-      "correct": 209,
-      "total": 225,
+      "accuracy": 0.988,
+      "correct": 596,
+      "total": 603,
       "top_confusions": {
-        "Personal Care": 4,
-        "Restaurants": 3,
-        "Travel": 3
+        "Healthcare": 6,
+        "Personal Care": 1
       }
     },
     "Subscription": {
-      "accuracy": 0.985,
-      "correct": 199,
-      "total": 202,
-      "top_confusions": {
-        "Education": 1,
-        "Personal Care": 1,
-        "Shopping": 1
-      }
+      "accuracy": 1.0,
+      "correct": 597,
+      "total": 597,
+      "top_confusions": {}
     },
     "Transfer": {
-      "accuracy": 0.995,
-      "correct": 216,
-      "total": 217,
-      "top_confusions": {
-        "Shopping": 1
-      }
+      "accuracy": 1.0,
+      "correct": 599,
+      "total": 599,
+      "top_confusions": {}
     },
     "Transportation": {
-      "accuracy": 0.992,
-      "correct": 246,
-      "total": 248,
-      "top_confusions": {
-        "Shopping": 1,
-        "Subscription": 1
-      }
+      "accuracy": 1.0,
+      "correct": 588,
+      "total": 588,
+      "top_confusions": {}
     },
     "Travel": {
       "accuracy": 1.0,
-      "correct": 191,
-      "total": 191,
+      "correct": 601,
+      "total": 601,
       "top_confusions": {}
     },
     "Utilities": {
       "accuracy": 1.0,
-      "correct": 211,
-      "total": 211,
+      "correct": 604,
+      "total": 604,
       "top_confusions": {}
     }
   }