We should load and save the dictionary if we want to do this right, but we are not building a production system. We just want an easy way to test data on the model. Since the dictionary is already in memory, we can store it as JSON. So we import json, check if the file exists, load it if it does, and create it if it does not.
Then we update the in memory dictionary with what we loaded, so new words keep getting added and token IDs stay consistent. We also make sure the update call writes changes back to disk right away. That is convenient, though it could slow things down if we update a lot.
After running, we confirm the file shows up, the dictionary is saved, and padding looks correct.
On this page of the site you can watch the video online Python Loading JSON Files with a duration of hours minute second in good quality, which was uploaded by the user Stephen Blum 15 May 2026, share the link with friends and acquaintances, this video has already been watched 76 times on youtube and it was liked by 3 viewers. Enjoy your viewing!