Tweet Classification into multiple Artificial Intelligence related Categories
This respository contains a bert model finetuned to classify a tweet/linkedin post. The model is trained as a multi label predictor, which will return multiple categories for each input. Model is trained on following 8 categories :
'AI News', 'AI Tools', 'AI Research', 'AI Models', 'AI Usecases','AI Open Source', 'Podcasts/Talks/Events', 'AI Opinions','Non AI'
Usage:
model = AutoModelForSequenceClassification.from_pretrained(
model_path,
)
tokenizer = AutoTokenizer.from_pretrained( model_path)
from transformers import pipeline
classify_tweet = pipeline("text-classification", model=model,tokenizer = tokenizer,return_all_scores=True)
output = classify_tweet (your text)
Labels used:
- LABEL_0 = 'AI News'
- LABEL_1 = 'AI Tools'
- LABEL_2 = 'AI Research'
- LABEL_3 = 'AI Models'
- LABEL_4 = 'AI Usecases'
- LABEL_5 = 'AI Open Source'
- LABEL_6 = 'Podcasts/Talks/Events'
- LABEL_7 = 'Opinions'
- LABEL_8 = 'Non AI'
- Downloads last month
- 1