AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
1min snip
Kate Park: Data Engines for Vision and Language
The Gradient: Perspectives on AI
Data Scaling Experiments Analysis
Plots showing data scaling experiments can reveal valuable insights. The correlation between data added over time and performance improvements is essential. While the relationship may not always be linear and could plateau, as long as performance keeps improving with reasonable effort, continuing the data track is beneficial. Once performance plateaus or additional data doesn't yield substantial gains, it's time to move on. The initial guess of the necessary data amount can be challenging, but constant model retraining can help refine this. Limited resources for training may result in overshooting the data collection. Balancing effort with tangible gains is crucial for data scaling improvements.
00:00
Transcript
Episode notes
Get the Snipd
podcast app
Unlock the knowledge in podcasts with the podcast player of the future.
AI-powered
Discover
highlights
Listen to the best highlights from the podcasts you love and dive into the full episode
Save any
moment
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Share
& Export
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Discover
highlights
Listen to the best highlights from the podcasts you love and dive into the full episode