Post

Replies

Boosts

Views

Activity

Comment on CreatML stop training
Thanks for your answer! There is more information: • training/validation: 10,000,000 • macOS: 15.3 • Xcode: 16.2 • Memory: 48GB I think it may be a problem of insufficient memory, because the amount of data is large. So if so, I want to know how to solve this problem (such as choosing to occupy hard disk space instead of memory).
Jan ’25
Comment on How to confirm whether CreatML is training
Thanks for your help. I found the right solution and trained my model correctly in about 24 hours. However,if you could help me with this question, it would be greatly appreciated. I noticed that my post-training model’s accuracy didn’t meet my expectations with 1,000,000 training and verification data. So, I increased the data to 10 times the original, resulting in 10,000,000 training and verification data.With the Boosted Tree algorithm,how long will it take to train this model on my machine?
Jan ’25
Comment on Code with Swift Assist
@zentention There are already hints of Swift Assist in Xcode 16.2 Beta (for instance, “Swift Assist” has been added to the change record, except for “Source Control,” indicating that all changes in Swift Assist will be recorded). However, it’s not yet available for use. Swift Assist will be released in Xcode 16.2 RC, and I’m eagerly anticipating its release.
Dec ’24
Comment on Check Room
Although I appreciate your response, there may be some misunderstandings.I hope to detect whether the user is in the room (not a specific room, but any room) or in a non-room environment (such as outdoors, airplanes).
Topic: Spatial Computing SubTopic: ARKit Tags:
Oct ’24
Comment on Anchor Wall
Thank you for your response. However, the issue we are discussing may not be a problem. My question arises from this point: https://developer.apple.com/documentation/visionos/placing-content-on-detected-planes
Topic: Spatial Computing SubTopic: ARKit Tags:
Oct ’24