few-shot learning(三)：Pretraining and Fine Tuning
优点：简单，准确度高。 数学基础 余弦相似度 softmax softmax会把大的值变大，小的值变小。 Few-shot Prediction Using Pretrained CNN 预训练一个大网络提取特征，预测时用到这个网络。可以使用多种方法训练（训练方法以及网络结构对结果会有影响），但是使用监督学习训练完，要把全连接层去掉。 few-shot分类方法 拿Query与μ做对比。
相关文章
- 1. pretraining+fine tuning
- 2. transfer learning and fine tuning
- 3. Fine-tuning
- 4. Caffe Knowledge about Softmax and fine tuning
- 5. Unsupervised Person Re-identification: Clustering and Fine-tuning
- 6. fine-tuning
- 7. Unsupervised Person Re-identification: Clustering and Fine-tuning
- 8. fine-tuning的二三事
- 9. TOMM2018_Unsupervised Person Re-identification: Clustering and Fine-tuning
- 10. caffe fine tuning
- 更多相关文章...
- • W3C RDF and OWL 活动 - W3C 教程
- • XSL-FO table-and-caption 对象 - XSL-FO 教程
- • RxJava操作符（七）Conditional and Boolean
- • RxJava操作符（三）Filtering
相关标签/搜索
每日一句
-
每一个你不满意的现在，都有一个你没有努力的曾经。
欢迎关注本站公众号,获取更多信息
相关文章
- 1. pretraining+fine tuning
- 2. transfer learning and fine tuning
- 3. Fine-tuning
- 4. Caffe Knowledge about Softmax and fine tuning
- 5. Unsupervised Person Re-identification: Clustering and Fine-tuning
- 6. fine-tuning
- 7. Unsupervised Person Re-identification: Clustering and Fine-tuning
- 8. fine-tuning的二三事
- 9. TOMM2018_Unsupervised Person Re-identification: Clustering and Fine-tuning
- 10. caffe fine tuning