TAUR-dev/rankalign-v6-gemma-2-2b-d0.15-e2-hc-b2d-dbl-all-fsx

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:2.6BQuant:BF16Ctx Length:8kPublished:Apr 6, 2026Architecture:Transformer Warm

The TAUR-dev/rankalign-v6-gemma-2-2b-d0.15-e2-hc-b2d-dbl-all-fsx is a 2.6 billion parameter Gemma-2-2b base model fine-tuned as part of the rankalign project. This model is specifically optimized for hypernym-concat-bananas-to-dogs-double-all tasks, focusing on identifying hierarchical relationships between concepts. It is designed for research and evaluation in semantic relation extraction, particularly for hypernymy, offering a specialized tool for linguistic analysis.

Loading preview...

Model Overview

This model, TAUR-dev/rankalign-v6-gemma-2-2b-d0.15-e2-hc-b2d-dbl-all-fsx, is a specialized fine-tuned checkpoint derived from the google/gemma-2-2b base model. It is part of the rankalign project, which focuses on aligning language models for specific semantic tasks.

Key Characteristics

  • Base Model: Google's Gemma-2-2b, a 2.6 billion parameter language model.
  • Fine-tuning Objective: Optimized for the hypernym-concat-bananas-to-dogs-double-all task, indicating a focus on identifying and processing hypernym relationships within a specific dataset.
  • Training Details: Fine-tuned for 2 epochs with a delta of 0.15, utilizing a preference loss weight of 1 and enforcing force-same-x during training.

Intended Use Cases

This model is primarily suited for:

  • Research in Semantic Relations: Ideal for academic or research applications requiring precise identification of hypernymy.
  • Linguistic Analysis: Can be used to explore and evaluate hierarchical semantic structures in text.
  • Comparative Evaluation: Provides a specific checkpoint for reproducibility and comparison within the rankalign project's evaluation framework, as demonstrated by the provided evaluation scripts for various hypernym tasks (e.g., hypernym-bananas, hypernym-dogs).