See this I'm not convinced about. Having tested a bunch of hugging face models for sentiment analysis, the results are very very poor. We see so many uses of te

See this I'm not convinced about. Having tested a bunch of hugging face models for sentiment analysis, the results are very very poor. We see so many uses of text generation, but if document summarization is so good then where are all the startups doing this for money? I agree that GPT et Al perform well on the benchmarks, but that's not the same thing at all.