AraBERT-based Relation Classifier (Is Relation SADAF1)

This model predicts whether a pair of Arabic arguments is in a semantic relation or not. It is built on top of AraBERT v0.2 and enriched with additional handcrafted features.

base_model: aubmindlab/bert-base-arabertv02 pipeline_tag: text-classification library_name: transformers tags:

  • arabic
  • relation-classification
  • argument-mining
  • binary-classification datasets:
  • Is Relation SADAF1 View here

model_details:

architecture:

- "[CLS] embedding of Arg1 [SEP] Arg2"
- "Contextual embedding of A_word"
- "One-hot features of Has_Almasdar (yes, no, null)"
- "Fusion layer + classification head"

input:

  • "Argument pair (two sentences)"
  • "Single word feature (A_word)"
  • "Almasdar feature (categorical)"

output:

  • "Relation (1)"
  • "No Relation (0)"

Open In Colab


Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for SADAF-IMAMU/2-is-relation

Finetuned
(4011)
this model