@pipcook/plugins-tensorflow-resnet-model-define
Description
This plugin defines a image classification model. This model is resnet. The model is built using python tensorflow.
Required Parameters
| Parameter |
Type |
Description |
Optional Parameters
| Parameter |
Type |
Default Value |
Description |
| loss |
string |
categorical_crossentropy |
the loss function to be defined. Look more
|
| metrics |
string[] |
[ 'accuracy' ] |
metrics used to evaluate model. Look more
|
| learningRate |
number |
0.001 |
base learning rate for Adam |
| decay |
number |
0.05 |
decay rate for Adam |
| freeze |
boolean |
false |
Whether to freeze first several layers |
| labelMap |
object |
null |
label map, only specified when recoverPath is provided |
| recoverPath |
string |
null |
Recover path of pretrained model |
Example
"modelDefine": {
"package": "@pipcook/plugins-tensorflow-resnet-model-define",
"params": {
"freeze": true
}
},