Post

I mean, the advantage of open source models is that if you think one is missing data then you can further train it with more data — that’s the reason for HF What they state as a concern, is more of an inconvenience