Added ukr-ru config

config.json

@@ -43,11 +43,14 @@
     "lr_scheduler": "",
     "lr_scheduler_params": {},
     "use_grad_scaler": false,
+    "cudnn_enable": true,
+    "cudnn_deterministic": false,
+    "cudnn_benchmark": true,
+    "training_seed": 54321,
     "model": "vits",
     "num_loader_workers": 8,
     "num_eval_loader_workers": 8,
     "use_noise_augment": false,
-    "use_language_weighted_sampler": true,
     "audio": {
         "fft_size": 1024,
         "win_length": 1024,
@@ -61,8 +64,8 @@
         "ref_level_db": 20,
         "do_sound_norm": false,
         "log_func": "np.log",
-        "do_trim_silence": false,
-        "trim_db": 20.0,
+        "do_trim_silence": true,
+        "trim_db": 45,
         "do_rms_norm": false,
         "db_level": null,
         "power": 1.5,
@@ -75,7 +78,7 @@
         "do_amp_to_db_mel": true,
         "pitch_fmax": 640.0,
         "pitch_fmin": 0.0,
-        "signal_norm": true,
+        "signal_norm": false,
         "min_level_db": -100,
         "symmetric_norm": true,
         "max_norm": 4.0,
@@ -97,7 +100,7 @@
         "eos": "<EOS>",
         "bos": "<BOS>",
         "blank": "<BLNK>",
-        "characters": "a\u0105bc\u0107de\u0119fghijkl\u0142mn\u0144o\u00f3pqrs\u015btuvwxyz\u017a\u017c\u0430\u0431\u0432\u0433\u0491\u0434\u0435\u0454\u0436\u0437\u0438\u0456\u0457\u0439\u043a\u043b\u043c\u043d\u043e\u043f\u0440\u0441\u0442\u0443\u0444\u0445\u0446\u0447\u0448\u0449\u044c\u044e\u044f",
+        "characters": "\u0430\u0431\u0432\u0433\u0491\u0434\u0435\u0454\u0436\u0437\u0438\u0456\u0457\u0439\u043a\u043b\u043c\u043d\u043e\u043f\u0440\u0441\u0442\u0443\u0444\u0445\u0446\u0447\u0448\u0449\u044c\u044e\u044f\u044d\u0451\u044b\u044a",
         "punctuations": "!'(),-.:;? ",
         "phonemes": null,
         "is_unique": true,
@@ -108,51 +111,61 @@
     "loss_masking": null,
     "sort_by_audio_len": true,
     "min_audio_len": 32768,
-    "max_audio_len": 160000,
+    "max_audio_len": 224000,
     "min_text_len": 1,
     "max_text_len": Infinity,
     "compute_f0": false,
     "compute_linear_spec": true,
-    "precompute_num_workers": 0,
+    "precompute_num_workers": 12,
     "start_by_longest": false,
     "datasets": [
         {
             "name": "mailabs",
-            "path": "",
+            "path": "./logs/uk_UK",
             "meta_file_train": "",
-            "ignored_speakers": null,
+            "ignored_speakers": [
+                "obruchov",
+                "shepel"
+            ],
             "language": "uk",
             "meta_file_val": "",
             "meta_file_attn_mask": ""
         },
         {
             "name": "mailabs",
-            "path": "",
+            "path": "./logs/ru_RU",
             "meta_file_train": "",
-            "ignored_speakers": null,
-            "language": "pl",
+            "ignored_speakers": [
+                "minaev",
+                "nikolaev"
+            ],
+            "language": "ru",
             "meta_file_val": "",
             "meta_file_attn_mask": ""
         }
     ],
     "test_sentences": [
         [
-            "\u041c\u0456\u0436 \u043d\u0438\u043c\u0438 \u0437\u043d\u043e\u0432 \u0437\u0430\u043a\u0440\u0443\u0442\u0438\u0432\u0441\u044f \u043f\u0430\u043d, \u0437\u0430\u043a\u0440\u0443\u0442\u0438\u0432\u0441\u044c \u043e\u0441\u0430\u0432\u0443\u043b\u0430.",
+            "\u0412\u0435\u0441\u0435\u043b\u043a\u0430, \u0442\u0430\u043a\u043e\u0436 \u0440\u0430\u0439\u0434\u0443\u0433\u0430 \u043e\u043f\u0442\u0438\u0447\u043d\u0435 \u044f\u0432\u0438\u0449\u0435 \u0432 \u0430\u0442\u043c\u043e\u0441\u0444\u0435\u0440\u0456, \u0449\u043e \u044f\u0432\u043b\u044f\u0454 \u0441\u043e\u0431\u043e\u044e \u043e\u0434\u043d\u0443, \u0434\u0432\u0456 \u0447\u0438 \u0434\u0435\u043a\u0456\u043b\u044c\u043a\u0430 \u0440\u0456\u0437\u043d\u043e\u043a\u043e\u043b\u044c\u043e\u0440\u043e\u0432\u0438\u0445 \u0434\u0443\u0433.",
             "sumska",
             null,
             "uk"
         ],
         [
-            "A mo\u017ce Wokulski ma g\u0142ow\u0119?",
-            "nina_brown",
+            "\u0420\u0430\u0434\u0443\u0433\u0430, \u0430\u0442\u043c\u043e\u0441\u0444\u0435\u0440\u043d\u043e\u0435, \u043e\u043f\u0442\u0438\u0447\u0435\u0441\u043a\u043e\u0435 \u0438 \u043c\u0435\u0442\u0435\u043e\u0440\u043e\u043b\u043e\u0433\u0438\u0447\u0435\u0441\u043a\u043e\u0435 \u044f\u0432\u043b\u0435\u043d\u0438\u0435, \u043d\u0430\u0431\u043b\u044e\u0434\u0430\u0435\u043c\u043e\u0435 \u043f\u0440\u0438 \u043e\u0441\u0432\u0435\u0449\u0435\u043d\u0438\u0438 \u044f\u0440\u043a\u0438\u043c \u0438\u0441\u0442\u043e\u0447\u043d\u0438\u043a\u043e\u043c \u0441\u0432\u0435\u0442\u0430.",
+            "hajdurova",
             null,
-            "pl"
+            "ru"
         ]
     ],
     "eval_split_max_size": null,
     "eval_split_size": 0.01,
+    "use_speaker_weighted_sampler": false,
+    "speaker_weighted_sampler_alpha": 1.0,
+    "use_language_weighted_sampler": true,
+    "language_weighted_sampler_alpha": 1.0,
     "model_args": {
-        "num_chars": 83,
+        "num_chars": 52,
         "out_channels": 513,
         "spec_segment_size": 32,
         "hidden_channels": 192,
@@ -168,7 +181,7 @@
         "kernel_size_flow": 5,
         "dilation_rate_flow": 1,
         "num_layers_flow": 4,
-        "resblock_type_decoder": "1",
+        "resblock_type_decoder": "2",
         "resblock_kernel_sizes_decoder": [
             3,
             7,
@@ -204,7 +217,7 @@
             4,
             4
         ],
-        "use_sdp": false,
+        "use_sdp": true,
         "noise_scale": 1.0,
         "inference_noise_scale": 0.667,
         "length_scale": 1,
@@ -222,7 +235,7 @@
         "d_vector_dim": 0,
         "detach_dp_input": true,
         "use_language_embedding": true,
-        "embedded_language_dim": 2,
+        "embedded_language_dim": 4,
         "num_languages": 2,
         "language_ids_file": null,
         "use_speaker_encoder_as_loss": false,
@@ -256,7 +269,7 @@
     "speaker_encoder_loss_alpha": 1.0,
     "return_wav": true,
     "r": 1,
-    "num_speakers": 5,
+    "num_speakers": 2,
     "use_speaker_embedding": true,
     "speakers_file": null,
     "speaker_embedding_channels": 256,