Change default temp

app.py

@@ -114,15 +114,15 @@ with gr.Blocks(title="😻 KaniTTS - Text to Speech", theme=gr.themes.Default())
 
             with gr.Accordion("Settings", open=False):
                 temp = gr.Slider(
-                    minimum=0.1, maximum=1.5, value=0.6, step=0.05,
+                    minimum=0.1, maximum=1.5, value=1.4, step=0.05,
                     label="Temp", 
                 )
                 top_p = gr.Slider(
-                    minimum=0.1, maximum=1.0, value=0.95, step=0.05,
+                    minimum=0.1, maximum=1.0, value=1.4, step=0.05,
                     label="Top P", 
                 )
                 rp = gr.Slider(
-                    minimum=1.0, maximum=2.0, value=1.1, step=0.05,
+                    minimum=1.0, maximum=2.0, value=1.4, step=0.05,
                     label="Repetition Penalty", 
                 )
                 max_tok = gr.Slider(
@@ -156,14 +156,14 @@ with gr.Blocks(title="😻 KaniTTS - Text to Speech", theme=gr.themes.Default())
     with gr.Row():
 
         examples = [
-            ["Anyway, um, so, um, tell me, tell me all about her. I mean, what's she like? Is she really, you know, pretty?", "male", 0.6, 0.95, 1.1, 1200],
-            ["No, that does not make you a failure. No, sweetie, no. It just, uh, it just means that you're having a tough time...", "male", 0.6, 0.95, 1.1, 1200],
-            ["I-- Oh, I am such an idiot sometimes. I'm so sorry. Um, I-I don't know where my head's at.", "male", 0.6, 0.95, 1.1, 1200],
-            ["Got it. $300,000. I can definitely help you get a very good price for your property by selecting a realtor.", "female", 0.6, 0.95, 1.1, 1200],
-            ["Holy fu- Oh my God! Don't you understand how dangerous it is, huh?", "male", 0.6, 0.95, 1.1, 1200],
-            ["You make my days brighter, and my wildest dreams feel like reality. How do you do that?", "female", 0.6, 0.95, 1.1, 1200],
-            ["Great, and just a couple quick questions so we can match you with the right buyer. Is your home address still 330 East Charleston Road?", "female", 0.6, 0.95, 1.1, 1200],
-            ["Oh, yeah. I mean did you want to get a quick snack together or maybe something before you go?", "female", 0.6, 0.95, 1.1, 1200],
+            ["Anyway, um, so, um, tell me, tell me all about her. I mean, what's she like? Is she really, you know, pretty?", "male", 1.4, 0.95, 1.1, 1200],
+            ["No, that does not make you a failure. No, sweetie, no. It just, uh, it just means that you're having a tough time...", "male", 1.4, 0.95, 1.1, 1200],
+            ["I-- Oh, I am such an idiot sometimes. I'm so sorry. Um, I-I don't know where my head's at.", "male", 1.4, 0.95, 1.1, 1200],
+            ["Got it. $300,000. I can definitely help you get a very good price for your property by selecting a realtor.", "female", 1.4, 0.95, 1.1, 1200],
+            ["Holy fu- Oh my God! Don't you understand how dangerous it is, huh?", "male", 1.4, 0.95, 1.1, 1200],
+            ["You make my days brighter, and my wildest dreams feel like reality. How do you do that?", "female", 1.4, 0.95, 1.1, 1200],
+            ["Great, and just a couple quick questions so we can match you with the right buyer. Is your home address still 330 East Charleston Road?", "female", 1.4, 0.95, 1.1, 1200],
+            ["Oh, yeah. I mean did you want to get a quick snack together or maybe something before you go?", "female", 1.4, 0.95, 1.1, 1200],
         ]
 
 

util.py

@@ -17,7 +17,7 @@ class Config:
     start_of_text: int = 1
     end_of_text: int = 2
     max_new_tokens: int = 1200
-    temperature: float = .6
+    temperature: float = 1.4
     top_p: float = .95
     repetition_penalty: float = 1.1
 
@@ -141,7 +141,7 @@ class KaniModel:
         # Load model with proper configuration
         self.model = AutoModelForCausalLM.from_pretrained(
             self.conf.model_name,
-            torch_dtype=torch.bfloat16,
+            dtype=torch.bfloat16,
             device_map=self.conf.device_map,
             token=token,
             trust_remote_code=True  # May be needed for some models
@@ -223,49 +223,3 @@ class KaniModel:
         point_3 = time.time()
         return audio, text, self.time_report(point_1, point_2, point_3)
 
-
-class Demo:
-    def __init__(self):
-        self.audio_dir = './audio_examples'
-        os.makedirs(self.audio_dir, exist_ok=True)
-        self.sentences = [
-            "You make my days brighter, and my wildest dreams feel like reality. How do you do that?",
-            "Anyway, um, so, um, tell me, tell me all about her. I mean, what's she like? Is she really, you know, pretty?",
-            "Great, and just a couple quick questions so we can match you with the right buyer. Is your home address still 330 East Charleston Road?",
-            "No, that does not make you a failure. No, sweetie, no. It just, uh, it just means that you're having a tough time...",
-            "Oh, yeah. I mean did you want to get a quick snack together or maybe something before you go?",
-            "I-- Oh, I am such an idiot sometimes. I'm so sorry. Um, I-I don't know where my head's at.",
-            "Got it. $300,000. I can definitely help you get a very good price for your property by selecting a realtor.",
-            "Holy fu- Oh my God! Don't you understand how dangerous it is, huh?"
-        ]
-        self.urls = [
-            'https://www.nineninesix.ai/examples/kani/1.wav',
-            'https://www.nineninesix.ai/examples/kani/2.wav',
-            'https://www.nineninesix.ai/examples/kani/5.wav',
-            'https://www.nineninesix.ai/examples/kani/6.wav',
-            'https://www.nineninesix.ai/examples/kani/3.wav',
-            'https://www.nineninesix.ai/examples/kani/7.wav',
-            'https://www.nineninesix.ai/examples/kani/4.wav',
-            'https://www.nineninesix.ai/examples/kani/8.wav'
-        ]
-    
-    def download_audio(self, url: str, filename: str):
-        filepath = os.path.join(self.audio_dir, filename)
-        if not os.path.exists(filepath):
-            r = requests.get(url)
-            r.raise_for_status()
-            with open(filepath, 'wb') as f:
-                f.write(r.content)
-        return filepath
-
-    def get_audio(self, filepath: str):
-        arr, _ = librosa.load(filepath, sr=22050)
-        return arr
-
-    def __call__(self):
-        examples = []
-        for idx, (sentence, url) in enumerate(zip(self.sentences, self.urls), start=1):
-            filename = f"{idx}.wav"
-            filepath = self.download_audio(url, filename)
-            examples.append([sentence, self.get_audio(filepath)])
-        return examples