Spaces:
Running
Running
sidebar
Browse files
app.py
CHANGED
@@ -109,28 +109,38 @@ if "messages" not in st.session_state:
|
|
109 |
st.session_state.messages = initial_prompt
|
110 |
|
111 |
|
112 |
-
st.title("Speech
|
113 |
reddit_thread = 'https://www.reddit.com/r/dataisbeautiful/comments/17413bq/oc_speech_diarization_app_that_transcribes_audio'
|
114 |
|
115 |
with st.sidebar:
|
116 |
st.markdown('''
|
117 |
# How to Use
|
118 |
|
119 |
-
1. Enter a youtube link
|
120 |
-
2. "Chat" with the
|
121 |
|
122 |
Example prompts:
|
123 |
- Which speaker spoke the most?
|
124 |
- What are important keywords in the transcript for SEO?
|
125 |
''')
|
126 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
127 |
st.divider()
|
128 |
|
129 |
st.markdown(f'''
|
130 |
# About
|
131 |
|
132 |
Given an audio file or a youtube link this app will
|
133 |
-
- [x] 1.
|
134 |
- [x] 2. Transcribe each audio segment using [OpenAi Whisper API](https://platform.openai.com/docs/guides/speech-to-text/quickstart)
|
135 |
- [x] 3. Set up an LLM chat with the transcript loaded into its knowledge database, so that a user can "talk" to the transcript of the audio file.
|
136 |
|
|
|
109 |
st.session_state.messages = initial_prompt
|
110 |
|
111 |
|
112 |
+
st.title("Speech-to-Chat")
|
113 |
reddit_thread = 'https://www.reddit.com/r/dataisbeautiful/comments/17413bq/oc_speech_diarization_app_that_transcribes_audio'
|
114 |
|
115 |
with st.sidebar:
|
116 |
st.markdown('''
|
117 |
# How to Use
|
118 |
|
119 |
+
1. Enter a youtube link.
|
120 |
+
2. "Chat" with the video.
|
121 |
|
122 |
Example prompts:
|
123 |
- Which speaker spoke the most?
|
124 |
- What are important keywords in the transcript for SEO?
|
125 |
''')
|
126 |
|
127 |
+
api_key_input = st.text_input(
|
128 |
+
"OpenAI API Key to lift request limits (Coming soon)",
|
129 |
+
disabled=True,
|
130 |
+
type="password",
|
131 |
+
placeholder="Paste your OpenAI API key here (sk-...)",
|
132 |
+
help="You can get your API key from https://platform.openai.com/account/api-keys.", # noqa: E501
|
133 |
+
value=os.environ.get("OPENAI_API_KEY", None)
|
134 |
+
or st.session_state.get("OPENAI_API_KEY", ""),
|
135 |
+
)
|
136 |
+
|
137 |
st.divider()
|
138 |
|
139 |
st.markdown(f'''
|
140 |
# About
|
141 |
|
142 |
Given an audio file or a youtube link this app will
|
143 |
+
- [x] 1. Partition the audio according to the identity of each speaker (diarization) using `pyannote` [HuggingFace Speaker Diarization api](https://huggingface.co/pyannote/speaker-diarization-3.0)
|
144 |
- [x] 2. Transcribe each audio segment using [OpenAi Whisper API](https://platform.openai.com/docs/guides/speech-to-text/quickstart)
|
145 |
- [x] 3. Set up an LLM chat with the transcript loaded into its knowledge database, so that a user can "talk" to the transcript of the audio file.
|
146 |
|