A greenhouse-shade trainer writes weights as 64-bit floats and the artifact balloons. Which official size lever should they use?
Select an answer to reveal the explanation.
Short Explanation
The greenhouse trainer wrote 64-bit floats and the artifact ballooned. Switch to float32 and the file shrinks. Changing the S3 class leaves the weights fat, Rekognition does not change dtypes, and quantization is off the exam.
Full Explanation
Changing numeric types, such as float32 instead of float64, is an official lever for model size. S3 storage class does not change the weight dtype. Rekognition is irrelevant, and quantization is out of scope.