QDyLoRA: Quantized Dynamic Low-Rank Adaptation for Efficient Large Language Model Tuning.
Hossein RajabzadehMojtaba ValipourTianshu ZhuMarzieh TahaeiHyock Ju KwonAli GhodsiBoxing ChenMehdi RezagholizadehPublished in: CoRR (2024)
Keyphrases
- language model
- low rank
- language modeling
- n gram
- convex optimization
- missing data
- retrieval model
- probabilistic model
- test collection
- matrix completion
- linear combination
- information retrieval
- high order
- matrix factorization
- low rank matrix
- query expansion
- singular value decomposition
- translation model
- rank minimization
- query terms
- high dimensional data
- mixture model
- semi supervised
- decision trees