import streamlit as st from pages.scoreboard import validate_email from pages.scoreboard import write_email st.title("About") st.markdown( """ # Ori Speech-To-Text Arena """ ) st.markdown("""## Motivation""") st.markdown(""" The Ori Speech-to-Text Arena was created to provide a transparent and unbiased evaluation platform for comparing various ASR models by pitting our proprietary models against industry-standard solutions in a blind testing environment. The Arena enables real-world performance assessment through diverse audio inputs and human evaluations, helping us understand how different models handle various accents, speech patterns, and acoustic conditions. This platform not only showcases the capabilities of our models but also provides valuable insights for ongoing development and optimization of our speech recognition technology. """) st.markdown( """## Arena """ ) st.markdown( """ * The Arena allows a user to record or choose a random audio, in which speech will be recognized by two randomly selected models. After listening to the audio, and evaluating the output from both the models, the user can vote on which transcription they prefer. Due to the risks of human bias and abuse, model names are revealed only after a vote is submitted.""" ) st.markdown( "## Scoreboard" ) st.markdown( """ * The Scoreboard shows the performance of the models in the Arena. The user can see the overall performance of the models, the model with the highest win rate, and the model with the most appearances. The user can also see the win rates of each model, as well as the appearance distribution of each model.""" ) st.markdown("## Model Description") st.markdown("""### Proprietary ASR models by Oriserve""") st.markdown(""" * Ori Prime - Foundational, large, and stable. * Ori Swift - Lighter and faster than Ori Prime. * Ori Apex - The top-performing model, fast and stable. * Ori Apex XT - Enhanced with more training, though slightly less stable than Ori Apex. """) st.markdown("### Third-party ASR services") st.markdown(""" * Deepgram - Deepgram Nova-2 API * Azure - Azure Speech Services API """) st.markdown("## Future Scope") st.markdown(""" * Add more languages for ASR. * Add more Open-Source and Third-party ASR models. * Upload audio for testing and evaluation. """) st.markdown( "## Contact Us" ) st.markdown( "To inquire about our speech-to-text models and APIs, you can submit your email using the form below." ) with st.form("login_form"): st.subheader("Please enter your email") email = st.text_input("Email") submit_button = st.form_submit_button("Submit Form") if submit_button: if not email: st.error("Please fill in all fields") else: if not validate_email(email): st.error("Please enter a valid email address") else: st.session_state.logged_in = True st.session_state.user_email = email write_email(st.session_state.user_email) st.success("Thanks for submitting your email, our team will be in touch with you shortly!")