Add training data

training_data.json

@@ -1,37 +1,37 @@
 {
   "train_losses": [
-    2100.769680175781,
-    1210.195811767578,
-    1176.4252709960938,
-    1143.586368408203,
-    1115.7994873046875,
-    1091.0681024169921,
-    1068.894476928711,
-    1061.0595739746093,
-    1048.5140423583985,
-    1030.019833984375,
-    1021.6400866699219,
-    1014.914428100586,
-    1012.0101983642578,
-    1008.9616821289062,
-    1007.0355267333985
+    2136.5024682617186,
+    1198.3308361816407,
+    1161.5707946777343,
+    1123.9894409179688,
+    1090.3879553222657,
+    1057.5052581787108,
+    1034.7834246826171,
+    1018.692387084961,
+    1008.2613201904297,
+    1000.2829901123047,
+    993.2367407226562,
+    987.9076995849609,
+    984.4344122314453,
+    982.5031854248047,
+    980.7457598876953
   ],
   "test_losses": [
-    1278.497021484375,
-    1241.6509338378905,
-    1210.9466491699218,
-    1211.009942626953,
-    1160.8113220214843,
-    1111.9551544189453,
-    1074.8253112792968,
-    1102.0312622070312,
-    1043.075909423828,
-    1020.9971008300781,
-    1005.1833709716797,
-    1002.2283630371094,
-    997.1159881591797,
-    993.3500823974609,
-    993.1695037841797
+    1273.5531799316407,
+    1226.4591247558594,
+    1197.41689453125,
+    1164.7423767089845,
+    1098.9119689941406,
+    1046.8178466796876,
+    1028.4834930419922,
+    1017.2713775634766,
+    1018.4531524658203,
+    1009.554721069336,
+    1002.5865447998046,
+    985.553515625,
+    986.9031860351563,
+    991.0865600585937,
+    990.1863067626953
   ],
   "config": {
     "epochs": 15,
@@ -47,14 +47,14 @@
       "initial_mi_beta": 1.0,
       "final_mi_beta": 1.0,
       "mi_beta_shape": "constant",
-      "initial_tc_beta": 10.0,
+      "initial_tc_beta": 0.0,
       "final_tc_beta": 10.0,
-      "tc_beta_shape": "constant",
+      "tc_beta_shape": "custom",
       "initial_dw_beta": 0.2,
       "final_dw_beta": 1.0,
       "dw_beta_shape": "custom",
       "warmup_epoch_ratio": 0.2,
-      "free_bits": 0.15,
+      "free_bits": 0.75,
       "focal_loss_alpha": 0.5,
       "focal_loss_gamma": 2.0
     },
@@ -62,9 +62,9 @@
       "initial_mi_beta": 1.0,
       "final_mi_beta": 1.0,
       "mi_beta_shape": "constant",
-      "initial_tc_beta": 10.0,
+      "initial_tc_beta": 0.0,
       "final_tc_beta": 10.0,
-      "tc_beta_shape": "constant",
+      "tc_beta_shape": "custom",
       "initial_dw_beta": 0.2,
       "final_dw_beta": 1.0,
       "dw_beta_shape": "custom",
@@ -73,7 +73,7 @@
     "regularization": {
       "encoder_dropout": 0.1,
       "decoder_dropout": 0.1,
-      "free_bits": 0.15,
+      "free_bits": 0.75,
       "focal_loss_alpha": 0.5,
       "focal_loss_gamma": 2.0
     },
@@ -85,9 +85,9 @@
       "best_epoch": null
     }
   },
-  "final_train_loss": 1007.0355267333985,
-  "final_test_loss": 993.1695037841797,
+  "final_train_loss": 980.7457598876953,
+  "final_test_loss": 990.1863067626953,
   "total_epochs": 15,
-  "best_train_loss": 1007.0355267333985,
-  "best_test_loss": 993.1695037841797
+  "best_train_loss": 980.7457598876953,
+  "best_test_loss": 985.553515625
 }