Fine-tuning LLMs
Fine-tuning large language models for NLP tasks using standard fine-tuning, LoRA, and quantization.
A collection of experiments fine-tuning large language models for downstream NLP tasks including sentiment analysis and text classification. Covers three adaptation strategies:
- Standard fine-tuning — full parameter updates
- LoRA (Low-Rank Adaptation) — efficient parameter-efficient tuning
- Quantization — reduced-precision inference for deployment