Add training data

training_data.json

@@ -1,43 +1,21 @@
 {
   "train_losses": [
-    5755.956953125,
-    4527.261198730469,
-    3494.703161621094,
-    2589.8849584960935,
-    1721.1179077148438,
-    762.3653802490235,
-    596.5937448120117,
-    549.3049771118164,
-    532.7015045166015,
-    527.6343267822266,
-    522.560433959961,
-    525.4734405517578,
-    532.2424160766601,
-    542.7450305175781,
-    549.4707604980468
+    6022.34626953125,
+    4383.866735839843,
+    2710.9047668457033
   ],
   "test_losses": [
-    4864.913232421875,
-    3544.6339721679688,
-    3383.009033203125,
-    3733.9931518554686,
-    3762.6799560546874,
-    4619.136267089843,
-    4848.683056640625,
-    5241.984057617187,
-    5930.204077148437,
-    6055.067846679687,
-    6544.0708984375,
-    7091.799609375,
-    7371.5255859375,
-    7603.633618164063,
-    8138.669360351562
+    5278.615673828125,
+    3384.1504150390624,
+    2245.631726074219
   ],
   "config": {
     "epochs": 15,
-    "batch_size": 32,
-    "learning_rate": 0.0005,
+    "batch_size": 1024,
+    "max_learning_rate": 0.001,
     "sequence_size": 32,
+    "shuffle_batches": true,
+    "shuffle_within_batch": true,
     "adaptive_weighting": {
       "initial_weight_emb": 1.5,
       "final_weight_emb": 0.0,
@@ -45,22 +23,34 @@
       "initial_weight_raw": 0.4,
       "final_weight_raw": 1.0,
       "weight_raw_shape": "linear",
-      "initial_kl_beta": 0.0001,
-      "final_kl_beta": 0.6,
-      "kl_beta_shape": "cosine",
-      "warmup_epoch_ratio": 0.4
+      "initial_mi_beta": 0.0,
+      "final_mi_beta": 0.0,
+      "mi_beta_shape": "constant",
+      "initial_tc_beta": 5.0,
+      "final_tc_beta": 5.0,
+      "tc_beta_shape": "constant",
+      "initial_dw_beta": 0.02,
+      "final_dw_beta": 0.3,
+      "dw_beta_shape": "custom",
+      "warmup_epoch_ratio": 0.2
     },
-    "total_correlation_beta_multiplier": 10.0,
     "free_bits": 0.15,
     "focal_loss_alpha": 0.75,
     "focal_loss_gamma": 2.0,
-    "dropout_rate": 0.2,
+    "dropout_rate": 0.1,
     "enable_dropout_on_latent": true,
-    "enable_dropout_on_decoder": true
+    "enable_dropout_on_decoder": true,
+    "early_stopping": {
+      "enabled": true,
+      "patience": 3,
+      "min_delta": 0.01,
+      "triggered": true,
+      "best_epoch": 3
+    }
   },
-  "final_train_loss": 549.4707604980468,
-  "final_test_loss": 8138.669360351562,
-  "total_epochs": 15,
-  "best_train_loss": 522.560433959961,
-  "best_test_loss": 3383.009033203125
+  "final_train_loss": 2710.9047668457033,
+  "final_test_loss": 2245.631726074219,
+  "total_epochs": 3,
+  "best_train_loss": 2710.9047668457033,
+  "best_test_loss": 2245.631726074219
 }