HiNER-original-xlm-roberta-large
| Property | Value |
|---|---|
| Model Type | Named Entity Recognition |
| Base Architecture | XLM-RoBERTa Large |
| Author | cfilt |
| Framework | PyTorch 1.9.1, Transformers 4.14.0 |
What is HiNER-original-xlm-roberta-large?
HiNER-original-xlm-roberta-large is a Named Entity Recognition model built on the XLM-RoBERTa large architecture, specifically trained on the HiNER-original dataset. This model represents a specialized implementation for identifying and classifying named entities in text, leveraging the multilingual capabilities of XLM-RoBERTa.
Implementation Details
The model was trained using carefully selected hyperparameters including a learning rate of 3e-05, with training conducted over 10 epochs. It utilizes the Adam optimizer with betas=(0.9,0.999) and epsilon=1e-08, combined with a linear learning rate scheduler. The training process employed batch sizes of 16 for training and 8 for evaluation.
- Trained from scratch on HiNER-original dataset
- Incorporates advanced transformer architecture
- Optimized using Adam optimizer
- Implements linear learning rate scheduling
Core Capabilities
- Named Entity Recognition in multilingual contexts
- Efficient processing with batch operations
- Optimized for both training and evaluation workflows
- Built on robust transformer architecture
Frequently Asked Questions
Q: What makes this model unique?
This model combines the powerful XLM-RoBERTa large architecture with specialized training on the HiNER-original dataset, making it particularly effective for named entity recognition tasks. Its carefully tuned hyperparameters and training approach ensure optimal performance.
Q: What are the recommended use cases?
The model is best suited for Named Entity Recognition tasks, particularly in scenarios requiring multilingual capability. It's designed to handle both training and inference with different batch sizes, making it adaptable to various computational resources.
