forked from mindspore-Ecosystem/mindspore
fix readme of densenet121 and config of inceptionv3
This commit is contained in:
parent
2c91bace34
commit
29d289ccb7
|
@ -246,7 +246,7 @@ You can modify the training behaviour through the various flags in the `train.py
|
|||
| Dataset | ImageNet |
|
||||
| epochs | 120 |
|
||||
| outputs | probability |
|
||||
| train performance | Top1:75.13%; Top5:92.57% |
|
||||
| accuracy | Top1:75.13%; Top5:92.57% |
|
||||
|
||||
### Training performance results
|
||||
|
||||
|
|
|
@ -42,7 +42,7 @@ config_gpu = edict({
|
|||
'is_save_on_master': 0,
|
||||
'dropout_keep_prob': 0.5,
|
||||
'has_bias': True,
|
||||
'amp_level': 'O3'
|
||||
'amp_level': 'O0'
|
||||
})
|
||||
|
||||
config_ascend = edict({
|
||||
|
|
Loading…
Reference in New Issue