BART Fine-tuned Keyphrase Extraction Model
| Property | Value |
|---|---|
| Author | aglazkova |
| Downloads | 291,968 |
| Base Architecture | BART-base |
| Paper | Research Paper |
| Training Datasets | Krapivin, Inspec, KPTimes, DUC-2001, PubMed, NamedKeys |
What is bart_finetuned_keyphrase_extraction?
This is a specialized model based on BART architecture, fine-tuned specifically for keyphrase generation tasks across multiple domains. It's designed to extract and generate relevant keyphrases from various text types, including scientific papers and news articles. The model has been trained on a diverse set of datasets to ensure robust cross-domain performance.
Implementation Details
The model was trained using carefully selected hyperparameters, including a learning rate of 4e-5 and the AdamW optimizer. Training was conducted over 6 epochs with a batch size of 8, optimizing for keyphrase generation accuracy.
- Built on the BART-base architecture
- Implements sequence-to-sequence learning for keyphrase generation
- Supports both scientific and news domain texts
- Utilizes transfer learning between domains
Core Capabilities
- Cross-domain keyphrase generation
- Support for multiple text types (scientific, biomedical, news)
- Efficient processing with PyTorch backend
- Seamless integration with Transformers library
Frequently Asked Questions
Q: What makes this model unique?
This model stands out for its cross-domain capabilities and extensive training on diverse datasets, making it particularly effective for both scientific and news-related content. Its optimization for transfer learning between domains sets it apart from single-domain models.
Q: What are the recommended use cases?
The model is ideal for automatic keyphrase extraction from research papers, news articles, and scientific texts. It's particularly useful for content categorization, document indexing, and automated metadata generation.
