NLP_Labelling / app.py
Abu1998's picture
Update app.py
a9998f7 verified
import gradio as gr
from huggingface_hub import InferenceClient
"""
For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
"""
client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
def respond(
message,
history: list[tuple[str, str]],
system_message,
max_tokens,
temperature,
top_p,
):
# Prepare the messages, starting with the system message
messages = [{"role": "system", "content": system_message}]
# Add the conversation history to the messages
for user_message, assistant_response in history:
if user_message:
messages.append({"role": "user", "content": user_message})
if assistant_response:
messages.append({"role": "assistant", "content": assistant_response})
# Add the current user message
messages.append({"role": "user", "content": message})
response = ""
# Stream the response from the model
for message in client.chat_completion(
messages,
max_tokens=max_tokens,
stream=True,
temperature=temperature,
top_p=top_p,
):
token = message.choices[0].delta.content
response += token
yield response
"""
For information on how to customize the ChatInterface, peruse the Gradio docs: https://www.gradio.app/docs/chatinterface
"""
demo = gr.ChatInterface(
respond,
additional_inputs=[
gr.Textbox(
value="""You are tasked with labeling text data based on both emotion temperature and text type categories. The final output must be a 13-character code that consists of the following structure:
1. Emotion Temperature Code (2 characters):
- If the emotion is purely Cold: Use CC
- If the emotion is purely Warm: Use WW
- If the emotion is purely Hot: Use HH
- If the emotion is a mix, use one of the following:
- Cold and Warm: Use CW
- Warm and Hot: Use WH
- Cold and Hot: Use CH
2. Text Type Codes (next 9 digits):
Assign a digit for each of the following categories based on the presence in the text. Use 0 for categories not applicable:
1: Toxic
2: Appreciation
3: Constructive Criticism
4: Genuine Questions
5: Advice/Suggestions
6: Requests
7: Spam
8: Off-Topic
9: Engagement Boosters
3. Special Categories (last 2 digits):
If the text is Neutral/General: Set the 10th digit to 1; otherwise, set it to 0.
If the text contains Hate: Set the last digit (11th) to 1; otherwise, set it to 0.
Example:
For the text "I love your videos but still something is missing":
- Emotion: Cold and Warm (CW)
- Types Detected: 2 (Appreciation), 3 (Constructive Criticism), 5 (Advice/Suggestions)
- Special Categories: Neutral/General (set the 10th digit to 1), no Hate
The output would be: CW02305000010
Output Format:
Always return a 13-character code following this structure.""",
label="each index of 13 digit have 0 to 9 , you need to extract the 13 digit number from the user input",
lines=10,
),
gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
gr.Slider(
minimum=0.1,
maximum=1.0,
value=0.95,
step=0.05,
label="Top-p (nucleus sampling)",
),
],
)
if __name__ == "__main__":
demo.launch()