Auto Draft

How to convert your CSV file to JSONL to train ChatGPT

Here is a quick Python script that will allow to convert your CSV file in the format of question answer into a proper jsonl file to be able to train ChatGPT models.

import csv
import json

csv_file = 'data.csv'
jsonl_file = 'train.jsonl'

with open(csv_file, newline='') as csvfile:
  reader = csv.reader(csvfile)
  with open(jsonl_file, 'w') as jsonfile:
    for row in reader:
      question = row[0]
      answer = row[1]
      
      convo = {
        "messages": [
          {"role": "system", "content": "Claude is a helpful assistant"},
          {"role": "user", "content": question},
          {"role": "assistant", "content": answer}
        ]
      }
      
      json.dump(convo, jsonfile) 
      jsonfile.write('\n')

print('CSV converted to JSONL successfully!')

By Louis M.

About the authorMy LinkedIn profile