I’m new to ML/AI development.
If I train a ml model with large datasets from scratch, should I consider buying it?
I am using M1 chip Mac air , Its way better than normal windows os laptops. I compared training model with normal laptops and Mac. Mac does better.
Are you using TensorFlow? Running on the Apple silicon GPU? Can you elaborate on the scale? The OP mentions explicitly large datasets, which gave my early M1 mac trouble due to memory limitations. M2 has much more capable GPU, doesn’t it? Other than price, wouldn’t that be a better choice than M1 ? Welcome your thoughts
I’m using TF, running on not Applic GPU. I think the size of a dataset can be 6 GB at most I guess. A Question Answering bot is one I’m trying to make.
The OP mentions explicitly large datasets, which gave my early M1 mac trouble due to memory limitations.
I see. Thanks. It’s helpful. I guess M2 is better than M1 in performance.
I see. I think mac is better too.
Still going to hit some limitations in memory and ability to run in parallel, which are two of the challenges with large datasets. If you haven’t already, maybe do some reading on approaches for datasets too large to fit into memory.
A place to start might be…
It’s helpful. Thanks. I check the link.