I don’t know much as i ain’t knowledgeable in ML and LLM, but i downloaded deepseek R1 7B model (which probably is the weights and architecture) from huggingface and it worked with no “data”. So i can say it’s reproducible. Now does that make it open source?
In capitalism, the wealthiests make the rule, not us.
All vloggers are paid by CPC, every Chinese you meet on the internet are bots. /s
While good quality subtitles are essential VLC can’t ensure that, it’s the responsibility of the production studio. AI subtitles on vlc are for those videos which doesn’t have any sub (which are a lot). The pushback shouldn’t be for vlc implementing AI, but production studios replacing translators or transcriber with AI (like crunchyroll tried last year).
Also while transcribing and subtitle editing is a labour intensive job, use of AI to help the editors shouldn’t be discouraged, it can increase their productivity by automating repeatative tasks so that they can focus on better quality.
They’re putting the children of the ISIS members in camps too. Even those who were born there. You think it is justified?