Epoch 50: 56.68%

README.md

@@ -5,7 +5,7 @@ tags:
 - cifar100
 - geometric-learning
 - fractal-encoding
-- trained
+- in-training
 - no-attention
 - no-cross-entropy
 datasets:
@@ -15,7 +15,7 @@ metrics:
 library_name: pytorch
 pipeline_tag: image-classification
 model-index:
-- name: geo-beatrix
+- name: geo-beatrix-step20-feats100
   results:
   - task:
       type: image-classification
@@ -30,13 +30,13 @@ model-index:
       verified: false
 ---
 
-# geo-beatrix
+# geo-beatrix-step20-feats100
 
 **Geometric Basin Classification for CIFAR-100**
 
-🎉 **Training Complete** 🎉
+🚧 **Training in Progress** 🚧
 
-Final Status: Epoch 200/200
+Current Status: Epoch 50/200
 
 ---
 
@@ -46,23 +46,23 @@ Final Status: Epoch 200/200
 |--------|-------|
 | **Best Test Accuracy** | **67.43%** |
 | **Best Epoch** | 195 |
-| **Current Train Accuracy** | 62.96% |
-| **Current Test Accuracy** | 67.13% |
-| **Current α (Cantor param)** | 0.4424 |
-| **Total Parameters** | 45,235,067 |
-| **Training Time** | 0:56:23 |
+| **Current Train Accuracy** | 61.27% |
+| **Current Test Accuracy** | 56.68% |
+| **Current α (Cantor param)** | 0.4217 |
+| **Total Parameters** | 46,338,737 |
+| **Training Time** | 0:13:18 |
 
 ### All Training Runs
 
 | Timestamp | Status | Best Epoch | Test Acc | Train Acc | α |
 |-----------|--------|------------|----------|-----------|---|
-| `20251010_014649` | ✅ | 195 | **67.43%** | 62.96% | 0.4424 |
+| `20251010_025042` | 🔄 | 195 | **67.43%** | 61.27% | 0.4217 |
 
 ### Comparison to State-of-the-Art
 
 | Model | Accuracy | Status |
 |-------|----------|--------|
-| **geo-beatrix (this model)** | **67.43%** | ✅ Complete |
+| **geo-beatrix (this model)** | **67.43%** | 🔄 Training |
 | vit-beatrix-dualstream | 66.0% | Vision Transformer + Cross-Entropy |
 | CLIP ViT-L/14 (zero-shot) | ~63-65% | 400M image-text pairs |
 | CLIP ViT-B/32 (zero-shot) | ~63.5% | Vision Transformer |
@@ -77,7 +77,7 @@ Final Status: Epoch 200/200
 - **Channels**: 64 → 128 → 256 → 512 → 1024
 - **Positional Encoding**: Devil's Staircase (Cantor function, 1883)
 - **PE Levels**: 20
-- **PE Features/Level**: 10
+- **PE Features/Level**: 100
 - **Classification**: Geometric Basin Compatibility (NO cross-entropy)
 - **Attention Mechanisms**: NONE
 
@@ -87,7 +87,7 @@ Final Status: Epoch 200/200
 
 ```json
 {
-  "model_name": "geo-beatrix",
+  "model_name": "geo-beatrix-step20-feats100",
   "model_type": "geometric_basin_classifier",
   "num_classes": 100,
   "batch_size": 256,
@@ -96,7 +96,7 @@ Final Status: Epoch 200/200
   "weight_decay": 0.05,
   "warmup_epochs": 10,
   "pe_levels": 20,
-  "pe_features_per_level": 10,
+  "pe_features_per_level": 100,
   "dropout": 0.1,
   "upload_every_n_epochs": 50,
   "alphamix": {
@@ -106,14 +106,14 @@ Final Status: Epoch 200/200
       0.7
     ],
     "spatial_ratio": 0.25,
-    "curriculum_start": 0.0,
-    "curriculum_end": 0.5
+    "curriculum_start": 0.05,
+    "curriculum_end": 0.4
   },
   "architecture": "ResNet-style with Devil's Staircase PE",
   "loss_function": "Geometric Basin Compatibility",
   "cross_entropy": false,
   "attention_mechanisms": false,
-  "timestamp": "20251010_014649"
+  "timestamp": "20251010_025042"
 }
 ```
 
@@ -127,7 +127,7 @@ 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/20251010_014649/
+├── weights/geo-beatrix-step20-feats100/20251010_025042/
 │   ├── model.pt                 (best from this training run)
 │   ├── model.safetensors        (best from this training run)
 │   ├── config.json
@@ -137,7 +137,7 @@ Final Status: Epoch 200/200
 │       ├── checkpoint_epoch_20.safetensors
 │       └── checkpoint_epoch_30.safetensors
 │       (snapshots every 50 epochs)
-└── runs/geo-beatrix/20251010_014649/
+└── runs/geo-beatrix-step20-feats100/20251010_025042/
     ├── events.out.tfevents.*    (TensorBoard logs)
     └── metrics.csv              (training metrics)
 ```
@@ -173,13 +173,13 @@ with open(info_path) as f:
 # Or download from specific training run
 model_path = hf_hub_download(
     repo_id="AbstractPhil/geo-beatrix",
-    filename="weights/geo-beatrix/20251010_014649/model.safetensors"
+    filename="weights/geo-beatrix-step20-feats100/20251010_025042/model.safetensors"
 )
 
 # Download specific epoch checkpoint
 epoch_checkpoint = hf_hub_download(
     repo_id="AbstractPhil/geo-beatrix",
-    filename="weights/geo-beatrix/20251010_014649/checkpoints/checkpoint_epoch_100.safetensors"
+    filename="weights/geo-beatrix-step20-feats100/20251010_025042/checkpoints/checkpoint_epoch_100.safetensors"
 )
 ```
 
@@ -189,26 +189,22 @@ epoch_checkpoint = hf_hub_download(
 
 ### Best Checkpoint
 - Epoch: 195
-- Train Acc: 58.21%
+- Train Acc: 61.27%
 - Test Acc: 67.43%
-- Alpha: 0.4424
-- Loss: 0.4848
+- Alpha: 0.4217
+- Loss: 0.0000
 
 ### Latest 5 Epochs
 
-- **Epoch 196**: Train 59.96%, Test 0.00%, α=0.4424, Loss=0.4921
-- **Epoch 197**: Train 60.25%, Test 0.00%, α=0.4424, Loss=0.4843
-- **Epoch 198**: Train 61.70%, Test 0.00%, α=0.4424, Loss=0.4852
-- **Epoch 199**: Train 63.14%, Test 0.00%, α=0.4424, Loss=0.4957
-- **Epoch 200**: Train 62.96%, Test 67.13%, α=0.4424, Loss=0.4965
+- **Epoch 46**: Train 59.31%, Test 0.00%, α=0.4166, Loss=1.1643
+- **Epoch 47**: Train 59.62%, Test 0.00%, α=0.4222, Loss=1.1372
+- **Epoch 48**: Train 60.62%, Test 0.00%, α=0.4214, Loss=1.1206
+- **Epoch 49**: Train 60.84%, Test 0.00%, α=0.4228, Loss=1.1334
+- **Epoch 50**: Train 61.27%, Test 56.68%, α=0.4217, Loss=1.1205
 
 ### Training Milestones
-- 🎯 **50% Accuracy** reached at epoch 40
-- 🎯 **60% Accuracy** reached at epoch 60
-- 🏆 **Beat vit-beatrix (66.0%)** at epoch 145
-- 🚀 **67% Accuracy** reached at epoch 175
-- 📊 **α ≥ 0.40** reached at epoch 9
-- 📊 **α ≥ 0.44** (near triadic equilibrium) at epoch 117
+- 🎯 **50% Accuracy** reached at epoch 35
+- 📊 **α ≥ 0.40** reached at epoch 10
 
 ---