Add training data

training_data.json

@@ -1,63 +1,82 @@
 {
   "train_losses": [
-    8840.943349609375,
-    5997.839943847656,
-    5144.925102539062,
-    5292.724506835937,
-    4626.860837402343,
-    4236.542509765625,
-    3929.224425048828,
-    3200.1126879882813,
-    2780.4872631835938,
-    2538.591198730469,
-    2381.423229980469,
-    2263.6071203613283,
-    2203.3307653808592,
-    2151.2380603027345,
-    2110.144202880859
+    158443.7097314453,
+    79525.31055175781,
+    55150.76360839844,
+    27544.720600585937,
+    21215.07262207031,
+    15201.550209960937,
+    17076.67155517578,
+    13131.717595214845,
+    11821.861442871093,
+    11532.774877929687,
+    12098.519575195312,
+    10800.846364746094,
+    11168.50675048828,
+    11143.719829101563,
+    10742.088754882812
   ],
   "test_losses": [
-    5255.883227539062,
-    4322.146154785156,
-    4260.568579101562,
-    4361.874780273438,
-    4510.580639648438,
-    5458.825817871094,
-    6068.390393066406,
-    6140.3800048828125,
-    6816.257507324219,
-    7403.943029785156,
-    8169.396142578125,
-    8850.116650390624,
-    8851.075305175782,
-    8749.574572753907,
-    9020.297473144532
+    38101.2255859375,
+    32535.043115234374,
+    12535.003198242188,
+    7116.39462890625,
+    8404.210815429688,
+    5529.595751953125,
+    5309.565930175781,
+    5373.602978515625,
+    5507.798205566406,
+    5121.186975097657,
+    4946.643518066407,
+    4945.580090332031,
+    5142.907373046875,
+    5019.152258300781,
+    4867.249340820313
   ],
   "config": {
     "epochs": 15,
-    "batch_size": 1024,
+    "batch_size": 32,
     "max_learning_rate": 0.001,
     "sequence_size": 32,
     "shuffle_batches": true,
     "shuffle_within_batch": true,
+    "vae_config": {
+      "latent_dim": 96,
+      "encoder_dropout": 0.1,
+      "decoder_dropout": 0.1,
+      "initial_mi_beta": 0.0,
+      "final_mi_beta": 0.0,
+      "mi_beta_shape": "constant",
+      "initial_tc_beta": 10.0,
+      "final_tc_beta": 10.0,
+      "tc_beta_shape": "constant",
+      "initial_dw_beta": 0.2,
+      "final_dw_beta": 1.0,
+      "dw_beta_shape": "linear",
+      "warmup_epoch_ratio": 0.2,
+      "free_bits": 0.15,
+      "focal_loss_alpha": 0.5,
+      "focal_loss_gamma": 2.0
+    },
     "adaptive_weighting": {
       "initial_mi_beta": 0.0,
       "final_mi_beta": 0.0,
       "mi_beta_shape": "constant",
-      "initial_tc_beta": 5.0,
-      "final_tc_beta": 5.0,
+      "initial_tc_beta": 10.0,
+      "final_tc_beta": 10.0,
       "tc_beta_shape": "constant",
-      "initial_dw_beta": 0.02,
+      "initial_dw_beta": 0.2,
       "final_dw_beta": 1.0,
       "dw_beta_shape": "linear",
       "warmup_epoch_ratio": 0.2
     },
-    "free_bits": 0.15,
-    "focal_loss_alpha": 0.75,
-    "focal_loss_gamma": 2.0,
-    "dropout_rate": 0.1,
-    "enable_dropout_on_latent": true,
-    "enable_dropout_on_decoder": true,
+    "regularization": {
+      "encoder_dropout": 0.1,
+      "decoder_dropout": 0.1,
+      "free_bits": 0.15,
+      "focal_loss_alpha": 0.5,
+      "focal_loss_gamma": 2.0
+    },
     "early_stopping": {
       "enabled": false,
       "patience": 3,
@@ -66,9 +85,9 @@
       "best_epoch": null
     }
   },
-  "final_train_loss": 2110.144202880859,
-  "final_test_loss": 9020.297473144532,
+  "final_train_loss": 10742.088754882812,
+  "final_test_loss": 4867.249340820313,
   "total_epochs": 15,
-  "best_train_loss": 2110.144202880859,
-  "best_test_loss": 4260.568579101562
+  "best_train_loss": 10742.088754882812,
+  "best_test_loss": 4867.249340820313
 }