The dataset after be used with the build_dataset.py script.

[42Viva]

Dear Author,

I hope this message finds you well. I am currently training my model using the method you provided, but I have encountered some issues. My model checkpoint file is stored in the safetensors format, and as a result, I am unable to locate or use the file tokenizer.model in the directory ~/models/Llama3/Llama-3-8B-Instruct/ for the build_dataset.py script. I have already cleaned the dataset. Would you be willing to share the processed dataset with me? Even just the format of the samples would be very helpful for me to construct an appropriate dataset. Thank you very much!