Training data format for Magicoder-OSS-Instruct-75K

[ise-uiuc]

Training data format for Magicoder-OSS-Instruct-75K

[UniverseFly]

Hi, here is the exact format we used when finetuning the model on OSS75K:

You are an exceptionally intelligent coding assistant that consistently delivers accurate and reliable responses to user instructions.

@@ Instruction
Write a solution to the following coding problem:
{problem}

@@ Response
{solution}

We haven't tested other templates ourselves, but we encourage anyone interested to explore them!

[VoiceBeer]

Thx! Appreciate it :>

[shatealaboxiaowang]

Hi, here is the exact format we used when finetuning the model on OSS75K:

You are an exceptionally intelligent coding assistant that consistently delivers accurate and reliable responses to user instructions.

@@ Instruction
Write a solution to the following coding problem:
{problem}

@@ Response
{solution}

We haven't tested other templates ourselves, but we encourage anyone interested to explore them!

Thx, is there a py script to convert the dataset Magicoder-OSS-Instruct-75K download from huggingface to the above format （instruction-response pairs）?

[shatealaboxiaowang]

Hi, here is the exact format we used when finetuning the model on OSS75K:

You are an exceptionally intelligent coding assistant that consistently delivers accurate and reliable responses to user instructions.

@@ Instruction
Write a solution to the following coding problem:
{problem}

@@ Response
{solution}

We haven't tested other templates ourselves, but we encourage anyone interested to explore them!

Thx, is there a py script to convert the dataset Magicoder-OSS-Instruct-75K download from huggingface to the above format （instruction-response pairs）?

@UniverseFly Thx, I have found the py script, yes it is preprocess_data.py