Datasets:

---

rasoul-nikbakht

/

TSpec-LLM

like 68

I am new to fine-tuning. As I know the dataset has to be in a specific format to fine tune a model. Like:
Instruction:
{}
input:
{}
output:
{}
The MD and docx files are huge. How can I turn them into this structured dataset?