Error in figure 1 #9
Description
Hi! Thank you for interesting idea!
If my understanding is correct, it seems that there are some error in Figure 1.
Since each model's parameter's delta is represented as (global diff) x (bit), it seems that the right side should be #models X (#parameters X 1bit + 16bit) instead of #parametersX (#models X 1bit + 16bit).
How do you think about this?