Can Perplexity Predict Fine-Tuning Performance? An Investigation of Tokenization Effects on Sequential Language Models for Nepali.
Nishant LuitelNirajan BekojuAnand Kumar SahSubarna ShakyaPublished in: CoRR (2024)
Keyphrases
- language model
- fine tuning
- n gram
- language modeling
- document retrieval
- probabilistic model
- word error rate
- retrieval model
- query expansion
- speech recognition
- viable alternative
- language modelling
- context sensitive
- test collection
- information retrieval
- named entities
- fine tuned
- pseudo relevance feedback
- statistical language models
- ad hoc information retrieval
- vector space model
- query terms
- relevance model
- smoothing methods
- machine learning
- query specific
- document ranking
- term dependencies
- out of vocabulary
- search engine