mr-checker/yt-strategic-intent-distilbert

Model Card: yt-strategic-intent-transformer

Model Details

• Model Name: yt-strategic-intent-transformer
• Task: Text Classification
• Framework: Hugging Face Transformers
• Library: PyTorch
• License: Apache 2.0

This model predicts strategic audience intent behind YouTube-related content. It answers the question: “What does the audience want?”

---

Intended Use

• Audience Insight: Identify strategic motivations behind content engagement.
• Content Strategy: Help creators and brands align delivery with audience intent.
• Market Research: Track evolving audience desires across categories.
• Business Intelligence: Map content to strategic drivers for decision-making.

---

Labels: Strategic Intent

• Build
• Learn
• Career
• Evaluate
• Awareness
• Decide
• Inspiration
• Predict

---

Training Data

• Source: yt-strategic-intent-9k
• Size: ~10,000 rows.
• Language: English (en).
• Preprocessing: Tokenization with DistilBERT tokenizer, balanced sampling across categories.

---

Evaluation

• Metrics: Accuracy, F1-score,
• Validation Strategy: Stratified train/validation split.

---

Example Usage

from transformers import pipeline

classifier = pipeline("text-classification", model="mr-checker/yt-strategic-intent-distilbert")

text = "How To Create AI UGC Ads (Full Tutorial)"
output = classifier(text)

print(output)

---

Limitations

• Supports only English-language inputs.
• Predictions depend on text quality (titles, transcripts, metadata).
• Does not yet incorporate multimodal signals (video thumbnails, audio tone).
• may misbehave sometimes on real data due to data quality