Add training dates and epoch-by-epoch training log

README.md

@@ -43,12 +43,15 @@ This repository contains the **pre-trained model artifacts** for [Nexus Shadow-Q
 
 ## 🏗️ Model Architecture
 
-| Model | Type | Parameters | Purpose |
-|:---|:---|:---|:---|
-| `predictor_v3.joblib` | XGBoost Ensemble | ~500 trees | Primary directional classifier |
-| `nexus_lstm_v3.pth` | Bi-LSTM | ~2M | Sequence pattern recognition |
-| `nexus_transformer_v2.pth` | Transformer | ~152M | Long-range dependency modeling |
-| `feature_scaler_v3.pkl` | StandardScaler | — | Feature normalization state |
+| Model | Type | Parameters | Trained | Purpose |
+|:---|:---|:---|:---|:---|
+| `predictor_v3.joblib` | XGBoost Ensemble | ~500 trees | 15 Feb 2026, 02:31 | Primary directional classifier |
+| `nexus_lstm_v3.pth` | Bi-LSTM | ~2M | 14 Feb 2026, 11:45 | Sequence pattern recognition |
+| `nexus_transformer_v2.pth` | Transformer (152M) | 5 epochs | 15 Feb 2026, 04:44 | Long-range dependency modeling |
+| `nexus_medium_transformer_v1.pth` | Transformer (Medium) | 5 epochs | 15 Feb 2026, 05:49 | Balanced capacity/speed |
+| `nexus_small_transformer_v1.pth` | Transformer (Small) | 10 epochs | 15 Feb 2026, 05:24 | Fast inference, high accuracy |
+| `nexus_transformer_pretrained.pth` | Pretrained base | — | 14 Feb 2026, 07:22 | Foundation weights |
+| `feature_scaler_v3.pkl` | StandardScaler | — | 15 Feb 2026, 02:31 | Feature normalization state |
 
 ### Supporting Models (16-Model Quant Panel)
 - **GARCH(1,1)** — Volatility regime detection
@@ -74,6 +77,60 @@ This repository contains the **pre-trained model artifacts** for [Nexus Shadow-Q
 
 ---
 
+## 🕐 Training Log
+
+<details>
+<summary><strong>📈 Small Transformer — 10 epochs (15 Feb 2026)</strong></summary>
+
+| Epoch | Accuracy | Timestamp |
+|:---|:---|:---|
+| 1 | 60.0% | 05:09 |
+| 2 | 69.7% | 05:10 |
+| 3 | 72.6% | 05:12 |
+| 4 | 74.5% | 05:14 |
+| 5 | 75.2% | 05:15 |
+| 6 | 76.0% | 05:17 |
+| 7 | 76.8% | 05:19 |
+| 8 | 76.8% | 05:20 |
+| 9 | 76.9% | 05:22 |
+| **10** | **76.9%** ✅ | **05:24** |
+
+</details>
+
+<details>
+<summary><strong>📈 Medium Transformer — 5 epochs (15 Feb 2026)</strong></summary>
+
+| Epoch | Accuracy | Timestamp |
+|:---|:---|:---|
+| 1 | 58.1% | 05:34 |
+| 2 | 69.8% | 05:37 |
+| 3 | 72.7% | 05:41 |
+| 4 | 74.8% | 05:45 |
+| **5** | **76.2%** ✅ | **05:49** |
+
+</details>
+
+<details>
+<summary><strong>📈 Nexus Transformer (152M) — 9 epochs (15 Feb 2026)</strong></summary>
+
+| Epoch | Accuracy | Timestamp |
+|:---|:---|:---|
+| 1 | 51.3% | 06:30 |
+| 2 | 52.4% | 06:51 |
+| 3 | 52.4% | 07:12 |
+| 4 | 53.1% | 07:32 |
+| 5 | 54.6% | 07:52 |
+| 6 | 55.3% | 08:13 |
+| 7 | 57.3% | 08:33 |
+| 8 | 58.1% | 08:54 |
+| **9** | **58.7%** ✅ | **09:14** |
+
+*Epoch 10 failed — weights from epoch 9 preserved.*
+
+</details>
+
+---
+
 ## ⚡ Quick Start
 
 ### Automatic (Recommended)
@@ -85,7 +142,7 @@ The Nexus Shadow-Quant app will **auto-pull** these models on first startup if n
 ### Manual
 ```bash
 pip install huggingface_hub
-huggingface-cli download Lukeexus/Predictor-Models --local-dir ./models
+huggingface-cli download Lukeed/Predictor-Models --local-dir ./models
 ```
 
 ---