app.py

from ArtistCoherencyModel import ArtistCoherencyModel
import streamlit as st
import pandas as pd

from LyricGeneratorModel import LyricGeneratorModel


@st.cache_resource
def get_artists():
    artists_df = pd.read_csv("artists.csv")
    return list(artists_df["name"])


@st.cache_resource
def get_evaluator_model():
    lyric_evaluator_model = None
    with st.spinner("Loading Evaluation Model"):
        lyric_evaluator_model = ArtistCoherencyModel.from_pretrained(
            "tjl223/artist-coherency-ensemble"
        )
    st.success("Finished Loading Evaluation Model")
    return lyric_evaluator_model


@st.cache_resource
def get_generator_model():
    lyric_generator_model = None
    with st.spinner("Loading Generator Model"):
        lyric_generator_model = LyricGeneratorModel(
            "tjl223/testllama2-qlora-lyric-generator-with-description"
        )
    st.success("Finished Loading Generator Model")
    return lyric_generator_model


lyric_evaluator_model = get_evaluator_model()
lyric_generator_model = get_generator_model()
artist_names_list = get_artists()

artist_name = st.selectbox("Artist", artist_names_list)
song_title = st.text_input("Song Title")
song_description = st.text_area("Song Description")
submit_button = st.button("Submit")

if submit_button:
    prompt = f"[Song Title] {song_title}\n[Song Artist] {artist_name}\n[Song Description] {song_description}"
    print(f"Prompt: {prompt}")
    st.write(prompt)
    lyrics = lyric_generator_model.generate_lyrics(prompt, 1000)
    print(f"Lyrics: {lyrics}")
    st.write(lyrics)
    score = lyric_evaluator_model.generate_artist_coherency_score(artist_name, lyrics)
    print(f"Score: {score}")
    st.write(f"Score: {score}")