kevinlu4588's picture
Training gradient ascent models and sanity checking car erasure
e225ed6
raw
history blame contribute delete
484 Bytes
import json
# Define the range and the common text description
start, end = 1, 1000
description = "a car"
# Open the metadata.jsonl file to write the JSON data
with open("metadata.jsonl", "w") as file:
# Iterate through the range and format each entry
for i in range(start, end + 1):
entry = {
"file_name": f"train/{i}.jpg",
"text": description
}
# Write each entry as a JSON line
file.write(json.dumps(entry) + "\n")