Automatically choose ModelStatistics output unit

[SniperTNT]

Implement #315

Automatically choose the unit for Total mult-adds, Input size, Forward/backward pass size, Params size, and Estimated Total Size

Output example 1:

==========================================================================================
Layer (type:depth-idx)                   Output Shape              Param #
==========================================================================================
small_model                              [16, 10]                  --
├─Linear: 1-1                            [16, 10]                  650
==========================================================================================
Total params: 650
Trainable params: 650
Non-trainable params: 0
Total mult-adds (K): 10.40
==========================================================================================
Input size (KB): 4.17
Forward/backward pass size (KB): 1.28
Params size (KB): 2.60
Estimated Total Size (KB): 8.05
==========================================================================================

Example 2:

.
.
.
Total params: 11,689,512
Trainable params: 11,689,512
Non-trainable params: 0
Total mult-adds (G): 181.41
==========================================================================================
Input size (MB): 60.21
Forward/backward pass size (GB): 3.97
Params size (MB): 46.76
Estimated Total Size (GB): 4.08
==========================================================================================