Automated NBA match commentary using OpenAI vision and TTS (with code!)
@skalskip9251 тыс. подписчиков
188 627 просмотров
10 комментариев
Комментарии из оригинального поста в X (Twitter)

@skalskip92
What is cool: Vision API noticed that it was a shot to win the game AND THE SERIES.
31 отметка «Нравится»

@thealexker
@NBA @OpenAI Nice remake! Original post and and my code
9 отметок «Нравится»

@JaxGeller
@NBA @OpenAI You can mix the crowd noise back in with Demucs! My narration didn't come out as well, but here's an example. Thanks for sharing.
8 отметок «Нравится»

@skalskip92
@NBA @OpenAI Awesome idea!

@Joybeanns
@NBA @OpenAI Why does it feel like its rapping 😅😅 great work.
3 отметки «Нравится»

@skalskip92
@NBA @OpenAI Not sure man! Hahaha we can make it rap 100% with a little bit of prompting haha
2 отметки «Нравится»

@shacrw_
@NBA @OpenAI v cool. small nitpicking. the words "as confetti flies..." came at 0:22, about ~5-8 seconds earlier than when the confetti actually appears. why is that? coz of batching?
2 отметки «Нравится»

@skalskip92
@NBA @OpenAI I uploaded all the images with the single API call. Not sure how it is processed. Maybe GPT have the access to all images at once.
1 отметка «Нравится»

@unk_data
@NBA @OpenAI Funny thing is it's literally a tutorial from OpenAI Cookbook.
2 отметки «Нравится»

@skalskip92
@NBA @OpenAI Please gimme link ;)