AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
AWS Developers Podcast
Episode 046 - Computer Vision on AWS with Francesco Pochetti – Part 2
Jul 29, 2022
Francesco Pochetti, Senior Machine Language Engineer at Bolt, dives deep into ML tools on AWS, such as NVIDIA Triton and TensorRT, improving processing time for Computer Vision. He also covers Amazon SageMaker, deploying ML models using Docker, and the importance of understanding the business problem when developing a machine learning model.
30:07
Episode guests
AI Summary
AI Chapters
Episode notes
Podcast summary created with Snipd AI
Quick takeaways
- Using image segmentation models in AWS SageMaker with TryTron and TensorRT drastically reduces latency and improves the naturalness of face blurring in photos.
- The four main steps of a machine learning pipeline are understanding the business problem, exploratory data analysis, modeling, and production, and utilizing tools like AWS SageMaker, Clarify, and SageMaker experiments is crucial for success.
Deep dives
Creating an Instagram-like filter for blurring faces in photos
The podcast episode discusses the process of creating an Instagram-like filter to automatically blur faces in photos of crowds. The idea came from a tweet by the CEO of Hugging Face, and the host decided to tackle the challenge by experimenting with off-the-shelf solutions for face detection and image processing. However, these solutions resulted in unnatural and rectangular blurs. To achieve a more natural effect, the host explored using a model for image segmentation, which classifies each pixel as either part of the face or the background. Training the model was easy with AWS SageMaker, and deployment became simpler with the integration of TryTron and TensorRT. The combined solution drastically reduced latency, making it 10 times faster. The progress in deployment and the use of Docker containers were also highlighted.
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