Update app.py

app.py

@@ -13,38 +13,43 @@ import commons
 import utils
 from data_utils import TextAudioLoader, TextAudioCollate, TextAudioSpeakerLoader, TextAudioSpeakerCollate
 from models import SynthesizerTrn
-from text.symbols import symbols_gra # import symbols graphemes
+from text.symbols import symbols as symbols_default # import symbols graphemes
 from text.symbols_ft import symbols_ft # import symbols finetuned model
 from scipy.io.wavfile import write
 from text import cleaners
 
-symbols = symbols_gra
-_symbol_to_id = {s: i for i, s in enumerate(symbols)}
-_id_to_symbol = {i: s for i, s in enumerate(symbols)}
+model_configs = {
+    "Graphemes_finetuned": {
+        "path": "fr_wa_graphemes/G_30000.pth",
+        "symbols": symbols_ft
+    },
+    "Graphemes": {
+        "path": "wa_graphemes/G_258000.pth",
+        "symbols": symbols_default
+    }
+}
+
+# Global variables 
+net_g = None
+symbols = []
+_symbol_to_id = {}
+_id_to_symbol = {}
 
 def text_to_sequence(text, cleaner_names):
-  '''Converts a string of text to a sequence of IDs corresponding to the symbols in the text.
-    Args:
-      text: string to convert to a sequence
-      cleaner_names: names of the cleaner functions to run the text through
-    Returns:
-      List of integers corresponding to the symbols in the text
-  '''
-  sequence = []
-
-  clean_text = _clean_text(text, cleaner_names)
-  for symbol in clean_text:
-    symbol_id = _symbol_to_id[symbol]
-    sequence += [symbol_id]
-  return sequence
+    sequence = []
+    clean_text = _clean_text(text, cleaner_names)
+    for symbol in clean_text:
+        symbol_id = _symbol_to_id[symbol]
+        sequence += [symbol_id]
+    return sequence
 
 def _clean_text(text, cleaner_names):
-  for name in cleaner_names:
-    cleaner = getattr(cleaners, name)
-    if not cleaner:
-      raise Exception('Unknown cleaner: %s' % name)
-    text = cleaner(text)
-  return text
+    for name in cleaner_names:
+        cleaner = getattr(cleaners, name)
+        if not cleaner:
+            raise Exception('Unknown cleaner: %s' % name)
+        text = cleaner(text)
+    return text
 
 def get_text(text, hps):
     text_norm = text_to_sequence(text, hps.data.text_cleaners)
@@ -53,31 +58,25 @@ def get_text(text, hps):
     text_norm = torch.LongTensor(text_norm)
     return text_norm
 
-def load_model(model_path, hps):
+def load_model_and_symbols(tab_name):
+    global net_g, symbols, _symbol_to_id, _id_to_symbol
+    model_config = model_configs[tab_name]
+    symbols = model_config["symbols"]
+    _symbol_to_id = {s: i for i, s in enumerate(symbols)}
+    _id_to_symbol = {i: s for i, s in enumerate(symbols)}
+
     net_g = SynthesizerTrn(
-        len(symbols), # change here
+        len(symbols),
         hps.data.filter_length // 2 + 1,
         hps.train.segment_size // hps.data.hop_length,
         n_speakers=hps.data.n_speakers,
         **hps.model)
     _ = net_g.eval()
-    _ = utils.load_checkpoint(model_path, net_g, None)
-    return net_g
-
-hps = utils.get_hparams_from_file("wa_graphemes/config.json")
-#hps = utils.get_hparams_from_file("wa_graphemes/config.json")
-
-model_paths = {
-    "Graphemes": "wa_graphemes/G_258000.pth"
-}
-#"Graphemes_ft": "fr_wa_finetune/G_198000.pth"
-# Load the model
-net_g = load_model(model_paths["Graphemes"], hps)
+    _ = utils.load_checkpoint(model_config["path"], net_g, None)
 
 def tts(text, speaker_id, tab_name):
-    global net_g
-    net_g = load_model(model_paths[tab_name], hps)
-    sid = torch.LongTensor([speaker_id])  # speaker ID
+    load_model_and_symbols(tab_name)
+    sid = torch.LongTensor([speaker_id])  # speaker identity
     stn_tst = get_text(text, hps)
 
     with torch.no_grad():
@@ -90,13 +89,15 @@ def tts(text, speaker_id, tab_name):
 def create_tab(tab_name):
     with gr.TabItem(tab_name):
         gr.Markdown(f"### {tab_name} TTS Model")
-        tts_input1 = gr.TextArea(label="Text in Walloon on graphemes", value="")
+        tts_input1 = gr.TextArea(label="Text in Walloon on IPA phonemes", value="")
         tts_input2 = gr.Dropdown(label="Speaker", choices=["Male", "Female"], type="index", value="Male")
         tts_submit = gr.Button("Generate", variant="primary")
         tts_output1 = gr.Textbox(label="Message")
         tts_output2 = gr.Audio(label="Output")
         tts_submit.click(lambda text, speaker_id: tts(text, speaker_id, tab_name), [tts_input1, tts_input2], [tts_output1, tts_output2])
 
+hps = utils.get_hparams_from_file("configs/vctk_base.json")
+
 app = gr.Blocks()
 with app:
     gr.Markdown(