I indexed 669 GB of my GoPro videos using my M1 Max computer and local ML models

Article summary

The author indexed 669 GB of GoPro videos using their M1 Max computer and local ML models to find interesting moments from their cycling journey. The project allowed them to search for specific moments and send the best clips to their DaVinci Resolve timeline. The author's project used open-source ML models and was able to index 628 videos with a total duration of 15 hours and 13 minutes. The project's details and metrics are available in the author's blog post.

Main themes

• Video indexing
• ML models
• Local processing
• GoPro videos
• Content search
• AI applications

What commenters say

• The project's ability to index and search large amounts of video data locally has potential applications beyond the author's cycling journey.
• Some commenters are interested in using similar technology to index and search explicit content, but this may require specialized models or techniques.
• The M1 Max computer's performance is compared to other CPUs, with some commenters noting its advantages in terms of memory bandwidth and AI workload handling.
• Dedicated GPUs may perform better than the M1 Max for AI workloads, despite the M1 Max's high memory bandwidth.
• The project's use of open-source ML models and local processing is seen as a positive aspect by some commenters, who prefer to have control over their data and workflow.
• The technology has potential applications in video editing and content creation, and some commenters are interested in incorporating it into their workflow.