Template Designs
This category contains generic annotation templates without specific paper references. Copy one and change it for your task.
Copy any template and run it in Potato. New to the tool? Start with the quick-start guide.
Subcategories
| Design |
Description |
Reference |
| best-worst-scaling |
MaxDiff annotation where annotators select the best and worst items from a set for relative comparison |
Template |
| pairwise-preference |
Compare two items and select the preferred one |
Template |
| ranking-task |
Drag-and-drop ranking interface to order items from best to worst |
Template |
| Design |
Description |
Reference |
| image-classification |
Multi-class image classification with thumbnail preview and zoom controls |
Template |
| image-segmentation |
Draw polygon masks around objects for semantic segmentation tasks |
Template |
| object-detection |
Draw bounding boxes around objects for object detection model training |
Template |
| Design |
Description |
Reference |
| likert-scale-survey |
Multi-question survey using Likert scales to measure agreement, satisfaction, or frequency |
Template |
| survey-feedback |
Multi-question survey with Likert scales, text fields, and multiple choice |
Template |
| Design |
Description |
Reference |
| dialogue-act-labeling |
Classify utterances in conversations by their communicative function (question, statement, request, etc.) |
Template |
| fact-verification |
Verify claims as supported, refuted, or not enough information based on provided evidence |
Template |
| hate-speech-detection |
Identify and categorize hate speech, offensive language, and toxic content in text |
Template |
| intent-classification |
Classify user utterances into intents for chatbot and virtual assistant training |
Template |
| named-entity-recognition |
Span-based entity labeling for identifying people, organizations, locations, and more |
Template |
| reading-comprehension |
Evaluate question-answer pairs for reading comprehension by verifying answers and rating quality |
Template |
| relation-extraction |
Identify and classify relationships between entities in text (e.g., works-for, located-in, married-to) |
Template |
| sarcasm-detection |
Identify sarcastic statements and label their type and target in social media and conversational text |
Template |
| semantic-similarity |
Rate the semantic similarity between pairs of sentences on a continuous scale |
Template |
| sentiment-analysis |
Simple 3-way sentiment classification with radio buttons |
Template |
| toxicity-detection |
Multi-label toxicity classification with severity ratings for content moderation |
Template |
| triage-quick-annotation |
A reusable template for quick triage annotation |
N/A (Template) |
Quick Start
# Navigate to a template
cd templates/text/sentiment-analysis
# Copy to your project
cp -r . /path/to/your/project/
# Customize config.yaml for your needs
# Then run with Potato
potato start config.yaml
Customization Tips
- Replace the label list in
config.yaml with your categories
- Write your own guidelines into
annotation_instructions
- Match
sample-data.json to the shape of your own data
- Set
instances_per_annotator and annotation_per_instance for your annotator pool
Task Count
Total: 20 template designs across 4 subcategories