GissaMittJobb

Charger

No, the old model does not have the training data. It only has "model weights". You can conceptualize those as the abstract rules that the old model learned when it read the training data. By design, they are not supposed to memorize their training data.

I expressed myself poorly, this is what I meant - it has the "essence" of the training data, but of course not the verbatim training data.

To outperform the old model, the new model needs more than what the old model learned. It needs primary sources, ie the training data itself. Which is going to be deleted.

I wonder how valuable in relative terms the old training data is to the process, compared to just the new training data. I can't answer it, but it would be interesting to know.

I think we're in agreement with each other? The old model has the old training data, and then you train a new one on that model with new training data, right?

I guess it depends on how important old data is when building upon new models, which I fully admit I don't know the answer to. As I understand it though, new models are not trained fully from scratch, but instead are a continuation of the older model trained with new techniques/new data.

To speculate, I guess not having the older data present in the new training stages might make the attributes of that data be less pronounced in the new output model.

Maybe they could cheat the system by trying to distill that data out of the older models and put that into the training data, but I guess the risk of model collapse is not-insignificant there

Again, limited understanding here, take everything I speculate with a grain of salt

They essentially still have the information in the weights, so I guess they won't fret too much over not having it in the original training data.

Heads up - your location is showing in this picture

I'm guessing it means "Blow", so Cocaine in this context. The cocaine is burning

I honestly think it's dumber than that - I think he's still salty about wind farms being put up outside his golf course in Scotland and is taking his petty revenge on the world

I've noticed this when the nights get hot here - having my feet outside the blankets and in the path of the fan helps a great deal with not overheating.

That's right. The art of correctly handling savedInstanceState is unfortunately not exactly well understood

Yeah, you can use the Epsilon garbage collector in Java for a no-op garbage collection strategy.

For short-lived programs that do not risk hitting any memory constraints, it makes a lot of sense - zero time wasted on cleanup during execution, then you just do it all at the end when killing the program, which you can do in constant time since you don't need to reason about what should remain or not.

Your phone is trying to keep your battery alive. The lower the specs of the phone, then the more aggressive the OS is.

No, apps closing between switches is not a matter of battery, it's a core feature of Android related to the management of RAM. Whenever the OS needs more available RAM, the OS will close a backgrounded app to make those resources available. This is why it happens more frequently on low-end devices - these generally ship with less RAM.

Some misguided vendors will limit background execution in incorrect ways in the name of saving battery, but the general thing with apps living in background is a story of RAM.

Interestingly enough, apps are supposed to be built to cope with being closed down due to lack of RAM and then be restored seamlessly, but this is an art that is uncommonly done correctly in the Android development space. The OS support is there, though.