Update app.py

app.py

@@ -35,13 +35,18 @@ def image_to_text(url, use_api=True):
 
 # LLM
 def generate_story(story_idea):
+    # template = """
+    #     You are a professional song writter;
+    #     Generate a song based on a simple narrative, the song should be no more than 100 words.
+    #     Song should be in Nepali language
+    #     CONTEXT: {story_idea}
+    #     STORY:
+    #     """
     template = """
-        You are a professional song writter;
-        Generate a song based on a simple narrative, the song should be no more than 100 words.
-        Song should be in Nepali language
-        CONTEXT: {story_idea}
-        STORY:
-        """
+            you are a song writer, write a song using following context:
+            {story_idea}. 
+            Song should not be more than 100 words. Write in Nepali language
+            """
     prompt = PromptTemplate(input_variables=["story_idea"], template=template)
 
     story_llm = LLMChain(llm=OpenAI(model_name='gpt-3.5-turbo-0301', temperature=1), prompt=prompt, verbose=True)
@@ -50,7 +55,7 @@ def generate_story(story_idea):
 
 # text to speech
 def text_to_speech(story):
-    API_URL = "https://api-inference.huggingface.co/models/openai/whisper-large-v2"
+    API_URL = "https://api-inference.huggingface.co/models/espnet/kan-bayashi_ljspeech_vits"
     headers = {"Authorization": f"Bearer {HUGGINGFACEHUB_API_TOKEN}"}
 
     payloads = {
@@ -86,7 +91,8 @@ def main():
             st.write(image_description)
 
         story = generate_story(story_idea=image_description)
-        story_starter_text = "Yo ho Radio Nepal, prastut xa sun nai parne katha: "
+        # story_starter_text = "Yo ho Radio Nepal, prastut xa sun nai parne katha: "
+        story_starter_text = ""
         story = story_starter_text + story
 
         # Display story text on FE