Camembert Base Xnli
C
Camembert Base Xnli
Developed by mtheo
Fine-tuned on the French portion of the XNLI dataset based on the Camembert-base model, supporting French zero-shot classification
Downloads 72
Release Time : 8/7/2023
Model Overview
This is a zero-shot classification model specifically designed for French text, based on the Camembert-base architecture and fine-tuned on the XNLI dataset. It can be used for natural language inference and text classification tasks.
Model Features
French Zero-Shot Classification
One of the few models supporting French zero-shot classification, capable of classifying new categories without fine-tuning.
Natural Language Inference Capability
Can calculate the probability of logical relationships (entailment/contradiction) between premises and hypotheses.
Efficient Fine-Tuning
Optimized fine-tuning on the French XNLI dataset, achieving 81.4% accuracy on the validation set.
Model Capabilities
French Text Classification
Zero-Shot Learning
Natural Language Inference
Premise-Hypothesis Relationship Judgment
Use Cases
Text Classification
News Topic Classification
Classify French news topics (e.g., sports/politics/science) without training.
Achieved 85.95% accuracy in sports topic classification in examples.
Content Moderation
Inappropriate Content Detection
Zero-shot detection by defining inappropriate content labels.
Featured Recommended AI Models