I’ve been running into this issue a few times lately, and now I’m out of options. Specifically I’m trying to download the google/gemma-3-4b-it model locally from the files and versions page: google/gemma-3-4b-it at main. All of the other download links work fine – they download to my ~/Downloads folder and I’m off to the races. But tokenizer.json is a different beast – in Chrome it opens a new tab where it is (apparently) trying to download the contents of the file. For other models I would just wait, then scrape the download. But for this model I can wait an eternity for nothing.
Is this a Chrome issue?
Hoping someone has experienced my pain and can provide me with a workaround!