y'all how do i make a dataset?

unslothai / unsloth

Finetune Llama 3.2, Mistral, Phi & Gemma LLMs 2-5x faster with 80% less memory

https://unsloth.ai

Apache License 2.0

16.43k stars 1.14k forks source link

y'all how do i make a dataset? #716

Open Pjn11111 opened 3 months ago

Pjn11111 commented 3 months ago

Is there any template that can guide me onto creating a dataset?. PS I have folders and folders, txt files and csv files. How do i convert them into a dataset??? And I wanna train this uncensored model of llama3 bartowski/Lexi-Llama-3-8B-Uncensored-GGUF/Lexi-Llama-3-8B-Uncensored-IQ4_XS.gguf How do i do it? Any help will be appreciated 🙏

danielhanchen commented 3 months ago

Converting raw text into a dataset sadly is a tough task itself. I would try some data augmentation tool - sadly I'm not an expert in this field, so my opinions are most likely wrong