随笔分类 -  Deep-learning

Framework, etc.
二分法
摘要：bisection method 阅读全文

posted @ 2022-08-18 17:07 cdekelon 阅读(36) 评论(0) 推荐(0)

Check compute capability and tf version
摘要：$ python -c "import tensorflow as tf; tf.test.is_gpu_available(); print('version :' + tf.__version__)" 阅读全文

posted @ 2022-05-24 18:39 cdekelon 阅读(41) 评论(0) 推荐(0)

netron
摘要：https://github.com/lutzroeder/netron https://github.com/lutzroeder/netron/releases/tag/v5.5.6 阅读全文

posted @ 2022-02-17 23:42 cdekelon 阅读(19) 评论(0) 推荐(0)

SOTA state-of-the-art model
摘要：SOTA model 也就是state-of-the-art model的意思 阅读全文

posted @ 2020-07-28 15:44 cdekelon 阅读(235) 评论(0) 推荐(0)

retinanet keras
摘要：https://github.com/fizyr/keras-retinanet 阅读全文

posted @ 2020-07-03 10:55 cdekelon 阅读(116) 评论(0) 推荐(0)

fp16 fp32 int8
摘要：https://developer.nvidia.com/blog/tensorrt-integration-speeds-tensorflow-inference/ 阅读全文

posted @ 2020-07-01 11:12 cdekelon 阅读(382) 评论(0) 推荐(0)

IOU
摘要：https://www.zybuluo.com/Team/note/1175689 阅读全文

posted @ 2020-02-12 17:00 cdekelon 阅读(90) 评论(0) 推荐(0)

coco , voc
摘要：https://blog.ddlee.cn/posts/7845fb62/ 阅读全文

posted @ 2020-02-10 17:56 cdekelon 阅读(73) 评论(0) 推荐(0)

Deep Learning for Generic Object Detection: A Survey
摘要：Deep Learning for Generic Object Detection: A Survey https://blog.csdn.net/AliceLeeHX/article/details/82726792 （Chinese） https://arxiv.org/pdf/1809.02 阅读全文

posted @ 2018-09-25 14:29 cdekelon 阅读(147) 评论(0) 推荐(0)

overfitting
摘要：Interpretation: You can see the cost decreasing. It shows that the parameters are being learned. However, you see that you could train the model even 阅读全文

posted @ 2018-03-21 16:25 cdekelon 阅读(96) 评论(0) 推荐(0)

caffe output argmax layer
摘要：http://blog.csdn.net/ws_20100/article/details/49184037 http://blog.csdn.net/langb2014/article/details/50988275 (good !) https://www.cnblogs.com/yinhey 阅读全文

posted @ 2018-02-26 11:17 cdekelon 阅读(1172) 评论(0) 推荐(0)

caffe python API
摘要：http://wentaoma.com/2016/08/10/caffe-python-common-api-reference/ 阅读全文

posted @ 2018-02-23 16:08 cdekelon 阅读(146) 评论(0) 推荐(0)

center loss
摘要：http://blog.csdn.net/u014380165/article/details/76946339 阅读全文

posted @ 2018-02-05 11:25 cdekelon 阅读(87) 评论(0) 推荐(0)

caffe solver
摘要：From http://blog.csdn.net/czp0322/article/details/52161759 solver.prototxt sovler的主要作用就是交替调用前向传导和反向传导 (forward & backward) 来更新神经网络的连接权值，从而达到最小化loss，实际 阅读全文

posted @ 2018-02-03 00:58 cdekelon 阅读(187) 评论(0) 推荐(0)

TP TN FP FN
摘要：内容都为转载. From https://blog.argcv.com/articles/1036.c From http://blog.csdn.net/Relocy/article/details/51453950 Case: 假如某个班级有男生80人,女生20人,共计100人.目标是找出所有女 阅读全文

posted @ 2018-01-16 11:34 cdekelon 阅读(545) 评论(0) 推荐(0)

tensor flow
摘要：白话深度学习与TensorFlow https://item.jd.com/12228460.html 第1章　机器学习是什么 2 1.1　聚类 4 1.2　回归 5 1.3　分类 8 1.4　综合应用 10 1.5　小结 14 第2章　深度学习是什么 15 2.1　神经网络是什么 15 2.1.1 阅读全文

posted @ 2018-01-09 15:37 cdekelon 阅读(283) 评论(0) 推荐(0)

Useful links or blogs
摘要：http://blog.csdn.net/qq_27923041/article/category/6668316 阅读全文

posted @ 2018-01-04 17:03 cdekelon 阅读(104) 评论(0) 推荐(0)

TensorFlow
摘要：Refer to https://www.jiqizhixin.com/articles/2017-01-02-7 https://www.leiphone.com/news/201702/T5e31Y2ZpeG1ZtaN.html 1. TensorFlow 是一个使用数据流图（data flow 阅读全文

posted @ 2018-01-02 16:37 cdekelon 阅读(114) 评论(0) 推荐(0)

Keras
摘要：https://www.jiqizhixin.com/articles/2017-01-02-7 http://blog.csdn.net/u012162613/article/details/45397033 https://keras.io/ 1. Keras is a high-level n 阅读全文

posted @ 2018-01-02 11:36 cdekelon 阅读(137) 评论(0) 推荐(0)

导航