mt5-base-finetuned-Spanish
| Property | Value |
|---|---|
| License | Apache 2.0 |
| Base Model | google/mt5-base |
| Training Dataset | wiki_lingua |
| Primary Task | Spanish Text Summarization |
What is mt5-base-finetuned-Spanish?
mt5-base-finetuned-Spanish is a specialized language model fine-tuned for Spanish abstractive summarization tasks. Built upon Google's MT5-base architecture, this model has been specifically optimized using the wiki_lingua dataset to generate concise Spanish summaries. The model demonstrates strong performance metrics, including a ROUGE-1 score of 28.11 and a BERTScore of 72.25.
Implementation Details
The model was trained using a carefully tuned configuration with Adam optimizer, utilizing a linear learning rate scheduler with warmup steps. Training specifications include:
- Learning rate: 0.0005
- Batch size: 32 (4 base with 8 gradient accumulation steps)
- Training epochs: 5
- Label smoothing factor: 0.1
Core Capabilities
- Abstractive text summarization in Spanish
- ROUGE-1: 28.11, ROUGE-2: 12.09, ROUGE-L: 24.62
- Average generation length: 18.73 tokens
- BERTScore: 72.25
Frequently Asked Questions
Q: What makes this model unique?
This model specializes in Spanish text summarization, leveraging the powerful MT5 architecture with specific optimizations for the Spanish language. Its performance metrics and specialized training make it particularly suitable for Spanish content summarization tasks.
Q: What are the recommended use cases?
The model is best suited for abstractive summarization of Spanish text, particularly in scenarios requiring concise content generation while maintaining semantic accuracy. It's ideal for applications in content automation, document summarization, and Spanish language processing systems.
