resnet50-提高训练速度?凯拉斯

时间:2018-09-21 01:45:17

标签: keras classification resnet

我该提高速度吗?我的意思是损失正在减少。发型。

Epoch 1/30
4998/4998 [==============================] - 307s 62ms/step - loss: 0.6861 - acc: 0.6347
Epoch 2/30
4998/4998 [==============================] - 316s 63ms/step - loss: 0.6751 - acc: 0.6387
Epoch 3/30
4998/4998 [==============================] - 357s 71ms/step - loss: 0.6676 - acc: 0.6387
Epoch 4/30
4998/4998 [==============================] - 376s 75ms/step - loss: 0.6625 - acc: 0.6387
Epoch 5/30
4998/4998 [==============================] - 354s 71ms/step - loss: 0.6592 - acc: 0.6387
Epoch 6/30
4998/4998 [==============================] - 345s 69ms/step - loss: 0.6571 - acc: 0.6387
Epoch 7/30
4998/4998 [==============================] - 349s 70ms/step - loss: 0.6559 - acc: 0.6387

模型架构: resnet50 (CNN with skip connections)

除了有1个FC,我有两个。然后我将softmax输出更改为Sigmoid进行二进制分类。

积极训练数据数量:1806 负数个培训数据:3192

对于每个示例([0、0、1、1,...]),我的输出都用1或0表示

批次= 40,数量= 30,但这没关系,因为损失停止了

0 个答案:

没有答案