TheBloke

/

Llama-2-13B-chat-GGML

like 697

I want to fine tune this model with my own dataset. I will be using autotrain-advanced and I want to know what should the dataset look like
Should it be in format for example

[INST]What is the capital city of France? [/INST] The capital city of France is Paris.

I have seen many dataset in different format.

My question what format should my dataset look like to fine tune this model