fix text_enc_hidden_dim on configs

IAHispano · Jul 13, 2024 · 31f9181 · 31f9181
1 parent 9aec49e
commit 31f9181
Show file tree

Hide file tree

Showing 9 changed files with 127 additions and 18 deletions.
diff --git a/rvc/configs/v1/32000.json b/rvc/configs/v1/32000.json
@@ -4,7 +4,10 @@
     "seed": 1234,
     "epochs": 20000,
     "learning_rate": 1e-4,
-    "betas": [0.8, 0.99],
+    "betas": [
+      0.8,
+      0.99
+    ],
     "eps": 1e-9,
     "batch_size": 4,
     "fp16_run": true,
@@ -35,13 +38,45 @@
     "kernel_size": 3,
     "p_dropout": 0,
     "resblock": "1",
-    "resblock_kernel_sizes": [3,7,11],
-    "resblock_dilation_sizes": [[1,3,5], [1,3,5], [1,3,5]],
-    "upsample_rates": [10,4,2,2,2],
+    "resblock_kernel_sizes": [
+      3,
+      7,
+      11
+    ],
+    "resblock_dilation_sizes": [
+      [
+        1,
+        3,
+        5
+      ],
+      [
+        1,
+        3,
+        5
+      ],
+      [
+        1,
+        3,
+        5
+      ]
+    ],
+    "upsample_rates": [
+      10,
+      4,
+      2,
+      2,
+      2
+    ],
     "upsample_initial_channel": 512,
-    "upsample_kernel_sizes": [16,16,4,4,4],
+    "upsample_kernel_sizes": [
+      16,
+      16,
+      4,
+      4,
+      4
+    ],
     "use_spectral_norm": false,
     "gin_channels": 256,
     "spk_embed_dim": 109
   }
-}
+}
diff --git a/rvc/configs/v1/40000.json b/rvc/configs/v1/40000.json
@@ -4,7 +4,10 @@
     "seed": 1234,
     "epochs": 20000,
     "learning_rate": 1e-4,
-    "betas": [0.8, 0.99],
+    "betas": [
+      0.8,
+      0.99
+    ],
     "eps": 1e-9,
     "batch_size": 4,
     "fp16_run": true,
@@ -35,13 +38,43 @@
     "kernel_size": 3,
     "p_dropout": 0,
     "resblock": "1",
-    "resblock_kernel_sizes": [3,7,11],
-    "resblock_dilation_sizes": [[1,3,5], [1,3,5], [1,3,5]],
-    "upsample_rates": [10,10,2,2],
+    "resblock_kernel_sizes": [
+      3,
+      7,
+      11
+    ],
+    "resblock_dilation_sizes": [
+      [
+        1,
+        3,
+        5
+      ],
+      [
+        1,
+        3,
+        5
+      ],
+      [
+        1,
+        3,
+        5
+      ]
+    ],
+    "upsample_rates": [
+      10,
+      10,
+      2,
+      2
+    ],
     "upsample_initial_channel": 512,
-    "upsample_kernel_sizes": [16,16,4,4],
+    "upsample_kernel_sizes": [
+      16,
+      16,
+      4,
+      4
+    ],
     "use_spectral_norm": false,
     "gin_channels": 256,
     "spk_embed_dim": 109
   }
-}
+}
diff --git a/rvc/configs/v1/48000.json b/rvc/configs/v1/48000.json
@@ -4,7 +4,10 @@
     "seed": 1234,
     "epochs": 20000,
     "learning_rate": 1e-4,
-    "betas": [0.8, 0.99],
+    "betas": [
+      0.8,
+      0.99
+    ],
     "eps": 1e-9,
     "batch_size": 4,
     "fp16_run": true,
@@ -35,13 +38,45 @@
     "kernel_size": 3,
     "p_dropout": 0,
     "resblock": "1",
-    "resblock_kernel_sizes": [3,7,11],
-    "resblock_dilation_sizes": [[1,3,5], [1,3,5], [1,3,5]],
-    "upsample_rates": [10,6,2,2,2],
+    "resblock_kernel_sizes": [
+      3,
+      7,
+      11
+    ],
+    "resblock_dilation_sizes": [
+      [
+        1,
+        3,
+        5
+      ],
+      [
+        1,
+        3,
+        5
+      ],
+      [
+        1,
+        3,
+        5
+      ]
+    ],
+    "upsample_rates": [
+      10,
+      6,
+      2,
+      2,
+      2
+    ],
     "upsample_initial_channel": 512,
-    "upsample_kernel_sizes": [16,16,4,4,4],
+    "upsample_kernel_sizes": [
+      16,
+      16,
+      4,
+      4,
+      4
+    ],
     "use_spectral_norm": false,
     "gin_channels": 256,
     "spk_embed_dim": 109
   }
-}
+}
diff --git a/rvc/configs/v2/bigvgan/32000.json b/rvc/configs/v2/bigvgan/32000.json
@@ -32,6 +32,7 @@
         "inter_channels": 192,
         "hidden_channels": 192,
         "filter_channels": 768,
+        "text_enc_hidden_dim": 768,
         "n_heads": 2,
         "n_layers": 6,
         "kernel_size": 3,

diff --git a/rvc/configs/v2/bigvgan/40000.json b/rvc/configs/v2/bigvgan/40000.json
@@ -33,6 +33,7 @@
         "inter_channels": 192,
         "hidden_channels": 192,
         "filter_channels": 768,
+        "text_enc_hidden_dim": 768,
         "n_heads": 2,
         "n_layers": 6,
         "kernel_size": 3,

diff --git a/rvc/configs/v2/bigvgan/48000.json b/rvc/configs/v2/bigvgan/48000.json
@@ -32,6 +32,7 @@
         "inter_channels": 192,
         "hidden_channels": 192,
         "filter_channels": 768,
+        "text_enc_hidden_dim": 768,
         "n_heads": 2,
         "n_layers": 6,
         "kernel_size": 3,

diff --git a/rvc/configs/v2/hifigan/32000.json b/rvc/configs/v2/hifigan/32000.json
@@ -32,6 +32,7 @@
         "inter_channels": 192,
         "hidden_channels": 192,
         "filter_channels": 768,
+        "text_enc_hidden_dim": 768,
         "n_heads": 2,
         "n_layers": 6,
         "kernel_size": 3,

diff --git a/rvc/configs/v2/hifigan/40000.json b/rvc/configs/v2/hifigan/40000.json
@@ -32,6 +32,7 @@
         "inter_channels": 192,
         "hidden_channels": 192,
         "filter_channels": 768,
+        "text_enc_hidden_dim": 768,
         "n_heads": 2,
         "n_layers": 6,
         "kernel_size": 3,

diff --git a/rvc/configs/v2/hifigan/48000.json b/rvc/configs/v2/hifigan/48000.json
@@ -32,6 +32,7 @@
         "inter_channels": 192,
         "hidden_channels": 192,
         "filter_channels": 768,
+        "text_enc_hidden_dim": 768,
         "n_heads": 2,
         "n_layers": 6,
         "kernel_size": 3,