Added E2, F5 model types.

#9

Author: niknah

F5TTS.py

@@ -17,7 +17,7 @@
 from comfy.utils import ProgressBar
 from cached_path import cached_path
 sys.path.append(Install.f5TTSPath)
-from model import DiT # noqa E402
+from model import DiT,UNetT # noqa E402
 from model.utils_infer import ( # noqa E402
     load_model,
     preprocess_ref_audio_text,
@@ -28,6 +28,7 @@
 
 class F5TTSCreate:
     voice_reg = re.compile(r"\{(\w+)\}")
+    model_types = ["F5", "E2"]
     tooltip_seed = "Seed. -1 = random"
 
     def is_voice_name(self, word):
@@ -54,7 +55,33 @@ def load_voice(ref_audio, ref_text):
         )
         return main_voice
 
-    def load_model(self):
+    def load_model(self, model):
+        models = {
+            "F5": self.load_f5_model,
+            "E2": self.load_e2_model,
+        }
+        return models[model]()
+
+    def get_vocab_file(self):
+        return os.path.join(
+            Install.f5TTSPath, "data/Emilia_ZH_EN_pinyin/vocab.txt"
+            )
+
+    def load_e2_model(self):
+        model_cls = UNetT
+        model_cfg = dict(dim=1024, depth=24, heads=16, ff_mult=4)
+        repo_name = "E2-TTS"
+        exp_name = "E2TTS_Base"
+        ckpt_step = 1200000
+        ckpt_file = str(cached_path(f"hf://SWivid/{repo_name}/{exp_name}/model_{ckpt_step}.safetensors")) # noqa E501
+        vocab_file = self.get_vocab_file()
+        ema_model = load_model(
+            model_cls, model_cfg,
+            ckpt_file, vocab_file
+            )
+        return ema_model
+
+    def load_f5_model(self):
         model_cls = DiT
         model_cfg = dict(
             dim=1024, depth=22, heads=16,
@@ -64,10 +91,11 @@ def load_model(self):
         exp_name = "F5TTS_Base"
         ckpt_step = 1200000
         ckpt_file = str(cached_path(f"hf://SWivid/{repo_name}/{exp_name}/model_{ckpt_step}.safetensors")) # noqa E501
-        vocab_file = os.path.join(
-            Install.f5TTSPath, "data/Emilia_ZH_EN_pinyin/vocab.txt"
+        vocab_file = self.get_vocab_file()
+        ema_model = load_model(
+            model_cls, model_cfg,
+            ckpt_file, vocab_file
             )
-        ema_model = load_model(model_cls, model_cfg, ckpt_file, vocab_file)
         return ema_model
 
     def generate_audio(self, voices, model_obj, chunks, seed):
@@ -117,8 +145,8 @@ def generate_audio(self, voices, model_obj, chunks, seed):
         os.unlink(wave_file.name)
         return audio
 
-    def create(self, voices, chunks, seed=-1):
-        model_obj = self.load_model()
+    def create(self, voices, chunks, seed=-1, model="F5"):
+        model_obj = self.load_model(model)
         return self.generate_audio(voices, model_obj, chunks, seed)
 
 
@@ -141,6 +169,7 @@ def INPUT_TYPES(s):
                     "default": 1, "min": -1,
                     "tooltip": F5TTSCreate.tooltip_seed,
                 }),
+                "model": (F5TTSCreate.model_types,),
             },
         }
 
@@ -174,7 +203,7 @@ def remove_wave_file(self):
                 print("F5TTS: Cannot remove? "+self.wave_file.name)
                 print(e)
 
-    def create(self, sample_audio, sample_text, speech, seed=-1):
+    def create(self, sample_audio, sample_text, speech, seed=-1, model="F5"):
         try:
             main_voice = self.load_voice_from_input(sample_audio, sample_text)
 
@@ -184,7 +213,7 @@ def create(self, sample_audio, sample_text, speech, seed=-1):
             chunks = f5ttsCreate.split_text(speech)
             voices['main'] = main_voice
 
-            audio = f5ttsCreate.create(voices, chunks, seed)
+            audio = f5ttsCreate.create(voices, chunks, seed, model)
         finally:
             self.remove_wave_file()
         return (audio, )
@@ -233,6 +262,7 @@ def INPUT_TYPES(s):
                     "default": 1, "min": -1,
                     "tooltip": F5TTSCreate.tooltip_seed,
                 }),
+                "model": (F5TTSCreate.model_types,),
             }
         }
 
@@ -289,7 +319,7 @@ def load_voices_from_files(self, sample, voice_names):
             voices[voice_name] = self.load_voice_from_file(sample_file)
         return voices
 
-    def create(self, sample, speech, seed=-1):
+    def create(self, sample, speech, seed=-1, model="F5"):
         # Install.check_install()
         main_voice = self.load_voice_from_file(sample)
 
@@ -309,7 +339,7 @@ def create(self, sample, speech, seed=-1):
             voices = self.load_voices_from_files(sample, voice_names)
             voices['main'] = main_voice
 
-            audio = f5ttsCreate.create(voices, chunks, seed)
+            audio = f5ttsCreate.create(voices, chunks, seed, model)
         return (audio, )
 
     @classmethod

pyproject.toml

@@ -1,7 +1,7 @@
 [project]
 name = "comfyui-f5-tts"
 description = "Text to speech with F5-TTS"
-version = "1.0.4"
+version = "1.0.5"
 license = {text = "MIT License"}
 
 [project.urls]