generate method can optionally accept a phonetics argument

Allows forcing the model so speak a certain way, for example: a bow or take a bow.

kokoro.py

@@ -131,15 +131,20 @@ def forward(model, tokens, ref_s, speed):
     asr = t_en @ pred_aln_trg.unsqueeze(0).to(device)
     return model.decoder(asr, F0_pred, N_pred, ref_s[:, :128]).squeeze().cpu().numpy()
 
-def generate(model, text, voicepack, speed=1):
-    ps = phonemize(text)
+def generate(model, text, voicepack, speed=1, phonetics=None):
+    if phonetics is not None:
+        ps = phonetics
+    else:
+        ps = phonemize(text)
+    
     tokens = tokenize(ps)
     if not tokens:
         return None
     elif len(tokens) > 510:
         tokens = tokens[:510]
         print('Truncated to 510 tokens')
+    
     ref_s = voicepack[len(tokens)]
     out = forward(model, tokens, ref_s, speed)
     ps = ''.join(next(k for k, v in VOCAB.items() if i == v) for i in tokens)
-    return out, ps
+    return out, ps