Multilingual LLM Fine-tuning & Adaptation
15 repos
Tools and model variants for adapting large language models across different languages using parameter-efficient fine-tuning (PEFT) techniques and safetensors format. The cluster centers on localized versions of Llama 2 chat models for Dutch, Portuguese, Spanish, French, Turkish, and Italian, demonstrating how modern LLMs are extended to support non-English languages while keeping training and deployment efficient. Developers exploring this area will find examples of language-specific model adaptation, efficient serialization patterns, and EU-focused model variants.
peft ·12
safetensors ·12
eu ·1
text-generation ·1
