Epoch 50: 39.68%

README.md

@@ -5,7 +5,7 @@ tags:
 - cifar100
 - geometric-learning
 - fractal-encoding
-- trained
+- in-training
 - no-attention
 - no-cross-entropy
 datasets:
@@ -34,9 +34,9 @@ model-index:
 
 **Geometric Basin Classification for CIFAR-100**
 
-🎉 **Training Complete** 🎉
+🚧 **Training in Progress** 🚧
 
-Final Status: Epoch 200/200
+Current Status: Epoch 50/200
 
 ---
 
@@ -46,11 +46,11 @@ Final Status: Epoch 200/200
 |--------|-------|
 | **Best Test Accuracy** | **56.12%** |
 | **Best Epoch** | 160 |
-| **Current Train Accuracy** | 53.80% |
-| **Current Test Accuracy** | 39.51% |
-| **Current α (Cantor param)** | 0.4377 |
+| **Current Train Accuracy** | 40.78% |
+| **Current Test Accuracy** | 39.68% |
+| **Current α (Cantor param)** | 0.4132 |
 | **Total Parameters** | 28,561,101 |
-| **Training Time** | 0:21:41 |
+| **Training Time** | 0:06:42 |
 
 ### All Training Runs
 
@@ -66,6 +66,7 @@ Final Status: Epoch 200/200
 | `20251011_023128` | ✅ | 160 | **56.12%** | 54.65% | 0.5005 |
 | `20251011_025919` | ✅ | 160 | **56.12%** | 57.80% | 0.4994 |
 | `20251011_032343` | ✅ | 160 | **56.12%** | 53.80% | 0.4377 |
+| `20251011_034748` | 🔄 | 160 | **56.12%** | 40.78% | 0.4132 |
 | `20251010_200842` | ✅ | 180 | **53.61%** | 67.53% | 0.4442 |
 | `20251010_185133` | ✅ | 200 | **52.97%** | 69.87% | 0.4452 |
 
@@ -73,7 +74,7 @@ Final Status: Epoch 200/200
 
 | Model | Accuracy | Status |
 |-------|----------|--------|
-| **geo-beatrix (this model)** | **56.12%** | ✅ Complete |
+| **geo-beatrix (this model)** | **56.12%** | 🔄 Training |
 | geo-beatrix (50M params) | 69.0% | Geometric Basin CONV architecture |
 
 🎯 **Current target**: Beat geo-beatrix (69.0%) - Currently -12.88%
@@ -90,7 +91,7 @@ Final Status: Epoch 200/200
 - **PE Features/Level**: 1000
 - **Classification**: Geometric Basin Compatibility (NO cross-entropy)
 - **Attention Mechanisms**: NONE
-- **Mixing**: Fractal (triadic multi-patch)
+- **Mixing**: Standard (single patch)
 
 ---
 
@@ -119,12 +120,12 @@ Final Status: Epoch 200/200
   },
   "alphamix": {
     "enabled": true,
-    "fractal_mode": true,
+    "fractal_mode": false,
     "range": [
       0.3,
       0.7
     ],
-    "spatial_ratio": 0.25,
+    "spatial_ratio": 0.1,
     "curriculum_start": 0.0,
     "curriculum_end": 0.5,
     "fractal_steps": [
@@ -141,7 +142,7 @@ Final Status: Epoch 200/200
   "loss_function": "Geometric Basin Compatibility",
   "cross_entropy": false,
   "attention_mechanisms": false,
-  "timestamp": "20251011_032343"
+  "timestamp": "20251011_034748"
 }
 ```
 
@@ -155,7 +156,7 @@ Final Status: Epoch 200/200
 ├── best_model_info.json     (which epoch/run this came from)
 ├── runs_history.json        (all training runs and their results)
 ├── README.md
-├── weights/geo-beatrix-resnet34-step20-feats1000/20251011_032343/
+├── weights/geo-beatrix-resnet34-step20-feats1000/20251011_034748/
 │   ├── model.pt                 (best from this training run)
 │   ├── model.safetensors        (best from this training run)
 │   ├── config.json
@@ -165,7 +166,7 @@ Final Status: Epoch 200/200
 │       ├── checkpoint_epoch_100.safetensors
 │       └── checkpoint_epoch_150.safetensors
 │       (snapshots every 10 epochs)
-└── runs/geo-beatrix-resnet34-step20-feats1000/20251011_032343/
+└── runs/geo-beatrix-resnet34-step20-feats1000/20251011_034748/
     ├── events.out.tfevents.*    (TensorBoard logs)
     └── metrics.csv              (training metrics)
 ```
@@ -201,13 +202,13 @@ with open(info_path) as f:
 # Or download from specific training run
 model_path = hf_hub_download(
     repo_id="AbstractPhil/geo-beatrix-resnet",
-    filename="weights/geo-beatrix-resnet34-step20-feats1000/20251011_032343/model.safetensors"
+    filename="weights/geo-beatrix-resnet34-step20-feats1000/20251011_034748/model.safetensors"
 )
 
 # Download specific epoch checkpoint
 epoch_checkpoint = hf_hub_download(
     repo_id="AbstractPhil/geo-beatrix-resnet",
-    filename="weights/geo-beatrix-resnet34-step20-feats1000/20251011_032343/checkpoints/checkpoint_epoch_100.safetensors"
+    filename="weights/geo-beatrix-resnet34-step20-feats1000/20251011_034748/checkpoints/checkpoint_epoch_100.safetensors"
 )
 ```
 
@@ -217,22 +218,21 @@ epoch_checkpoint = hf_hub_download(
 
 ### Best Checkpoint
 - Epoch: 160
-- Train Acc: 58.15%
-- Test Acc: 41.74%
-- Alpha: 0.4387
-- Loss: 0.8552
+- Train Acc: 40.78%
+- Test Acc: 56.12%
+- Alpha: 0.4132
+- Loss: 0.0000
 
 ### Latest 5 Epochs
 
-- **Epoch 196**: Train 60.15%, Test 0.00%, α=0.4377, Loss=0.8412
-- **Epoch 197**: Train 54.50%, Test 0.00%, α=0.4377, Loss=0.7728
-- **Epoch 198**: Train 54.60%, Test 0.00%, α=0.4377, Loss=0.7710
-- **Epoch 199**: Train 54.85%, Test 0.00%, α=0.4377, Loss=0.7651
-- **Epoch 200**: Train 53.80%, Test 39.51%, α=0.4377, Loss=0.7805
+- **Epoch 46**: Train 39.24%, Test 0.00%, α=0.4101, Loss=1.7538
+- **Epoch 47**: Train 39.59%, Test 0.00%, α=0.4090, Loss=1.8065
+- **Epoch 48**: Train 40.28%, Test 0.00%, α=0.4128, Loss=1.8005
+- **Epoch 49**: Train 40.06%, Test 0.00%, α=0.4142, Loss=1.7817
+- **Epoch 50**: Train 40.78%, Test 39.68%, α=0.4132, Loss=1.7571
 
 ### Training Milestones
-- 📊 **α ≥ 0.40** reached at epoch 40
-- 📊 **α ≥ 0.44** (near triadic equilibrium) at epoch 104
+- 📊 **α ≥ 0.40** reached at epoch 17
 
 ---
 
@@ -243,7 +243,7 @@ epoch_checkpoint = hf_hub_download(
 ✅ **Fractal positional encoding** (Cantor function from 1883)  
 ✅ **Geometric compatibility classification**  
 ✅ **ResNet34 backbone** (proven CNN architecture)
-✅ **Triadic fractal mixing** (base-3 aligned)
+
 
 ---