Model
NVIDIA Deep Learning Examples
EfficientNet-B0 backbone savedmodel (TensorFlow2, AMP, Imagenet)EfficientNet-B0 TensorFlow2 savedmodel without prediction heads
Copied!
1 Version
21.02.0Selected02/03/2022 2:34 AM UTC20.15 MBAccuracy: 0.7766500 EpochsBatch Size: 0GPU: A100 Copied!
21.02.0Selected
02/03/2022 2:34 AM UTC20.15 MBAccuracy: 0.7766500 EpochsBatch Size: 0GPU: A100
Copied!
architecture
| Key | Value |
|---|---|
| task | Classification |
| type | B0 |
performance
| Key | Value |
|---|---|
| accuracy | 0.7766 |
training
| Key | Value |
|---|---|
| LR_decay | cosine |
| global_batch_size | 16384 |
| training_precision | AMP |
| LR_init | 0.64 |
| dataset | ImageNet |
| epochs | 500 |