Skip to main content
. 2021 Apr 9;2021:5580914. doi: 10.1155/2021/5580914

Table 3.

The number of parameters in each EfficientNet and the recommended input size.

Models Trainable parameters (million) Input size
B0 ∼4.3 224 × 224
B1 ∼6.8 240 × 240
B2 ∼8.0 260 × 260
B3 ∼11.0 300 × 300
B4 ∼17.9 380 × 380
B5 ∼28.7 456 × 456
B6 ∼41.1 528 × 528