Lahad/wolof_only-dataset
Domaine:
natural language processingType de record:
dataset
Créateur:
Lah
Hôte:
Train a BERT, GPT or LLaMA model specific to Wolof
Create word embeddings (Word2Vec, FastText)
Fine-tune multilingual models (mBERT, XLM-RoBERTa)
Next Token Prediction
Masked Language Modeling (MLM)
Perplexity calculation
Automatic text generation in Wolof
Sentence completion