文章分类 -  Deep-learning

1 2 3 下一页

hooks来简单可视化卷积神经网络的特征提取
摘要:https://developer.huawei.com/consumer/cn/forum/topic/0202545529428830025?fid=23 阅读全文

posted @ 2021-09-22 23:07 cdekelon 阅读(24) 评论(0) 推荐(0) 编辑

python - 深度学习Loss Nan的原因
摘要:python - 深度学习Loss Nan的原因 阅读全文

posted @ 2021-08-26 10:54 cdekelon 阅读(38) 评论(0) 推荐(0) 编辑

train loss 不断下降,test loss 不断下降,说明网络正在学习
摘要:From: https://www.cnblogs.com/jfdwd/p/11239564.html 参考:https://blog.csdn.net/LIYUAN123ZHOUHUI/article/details/744539801 train loss 不断下降,test loss 不断下降 阅读全文

posted @ 2021-08-20 23:51 cdekelon 阅读(312) 评论(0) 推荐(0) 编辑

http://ethereon.github.io/netscope/#/editor 画network flow
摘要:画network flow 阅读全文

posted @ 2021-06-19 21:34 cdekelon 阅读(24) 评论(0) 推荐(0) 编辑

early-stopping
摘要:https://machinelearningmastery.com/how-to-stop-training-deep-neural-networks-at-the-right-time-using-early-stopping/ https://blog.csdn.net/df19900725/ 阅读全文

posted @ 2021-06-17 17:42 cdekelon 阅读(40) 评论(0) 推荐(0) 编辑

onnxruntime
摘要:https://www.onnxruntime.ai/docs/tutorials/tutorials/tf-get-started.html https://azure.microsoft.com/en-us/blog/onnx-runtime-integration-with-nvidia-te 阅读全文

posted @ 2021-06-07 23:23 cdekelon 阅读(260) 评论(0) 推荐(0) 编辑

Blogs
摘要:https://sh-tsang.medium.com/review-mwcnn-multi-level-wavelet-cnn-for-image-restoration-denoising-super-resolution-jpeg-6bdf48cd6e3d https://analyticsi 阅读全文

posted @ 2021-06-06 00:17 cdekelon 阅读(107) 评论(0) 推荐(0) 编辑

Zero-center by mean pixel rgb-->bgr keras_applications/imagenet_utils.py
摘要:https://github.com/keras-team/keras-applications/blob/master/keras_applications/imagenet_utils.py https://www.tutorialfor.com/questions-30132.htm http 阅读全文

posted @ 2021-02-10 15:45 cdekelon 阅读(63) 评论(0) 推荐(0) 编辑

https://petewarden.com/2017/12/14/how-many-images-do-you-need-to-train-a-neural-network/
摘要:https://petewarden.com/2017/12/14/how-many-images-do-you-need-to-train-a-neural-network/ This is what I believe in: “You need as much data as the numb 阅读全文

posted @ 2021-01-29 00:51 cdekelon 阅读(47) 评论(0) 推荐(0) 编辑

bbox is known, how to draw bbox in the image
摘要:import os import cv2 for i in sorted(os.listdir("./images_pad")): id = os.path.splitext(i)[0] with open("./labels_pad/" + id + ".txt", "r") as f: for 阅读全文

posted @ 2021-01-18 15:10 cdekelon 阅读(18) 评论(0) 推荐(0) 编辑

standford drone data
摘要:https://www.alexkras.com/convert-video-to-images-via-command-line/ https://github.com/JosephKJ/SDD-Utils/blob/master/annotate.py https://github.com/ro 阅读全文

posted @ 2020-09-11 01:14 cdekelon 阅读(100) 评论(0) 推荐(0) 编辑

amazon sagemaker
摘要:https://github.com/aws-samples/amazon-sagemaker-nvidia-ngc-examples/blob/master/TensorFlow_BYOC_for_training_from_NGC/tf-bring-your-container.ipynb ht 阅读全文

posted @ 2020-09-08 16:06 cdekelon 阅读(57) 评论(0) 推荐(0) 编辑

model parallel
摘要:https://mxnet.apache.org/versions/1.6/api/faq/model_parallel_lstm.html https://medium.com/@esaliya/model-parallelism-in-deep-learning-is-not-what-you- 阅读全文

posted @ 2020-08-05 15:51 cdekelon 阅读(82) 评论(0) 推荐(0) 编辑

Draw process flow 流程图
摘要:https://app.diagrams.net/ 阅读全文

posted @ 2020-07-30 12:12 cdekelon 阅读(155) 评论(0) 推荐(0) 编辑

用程序理解nms
摘要:转载 https://github.com/SnailTyan/deep-learning-tools/blob/master/nms.py #!/usr/bin/env python # _*_ coding: utf-8 _*_ import cv2 import numpy as np """ 阅读全文

posted @ 2020-07-14 16:43 cdekelon 阅读(188) 评论(0) 推荐(0) 编辑

0 1 2 轴
摘要:https://blog.csdn.net/u012762410/article/details/78912667 阅读全文

posted @ 2020-07-05 00:47 cdekelon 阅读(63) 评论(0) 推荐(0) 编辑

c++ trt engine
摘要://g++ inferPeopleNet.cpp `pkg-config --cflags --libs opencv4` -I /usr/local/cuda-10.2/include -I /usr/include/aarch64-linux-gnu/ -L /usr/lib/aarch64-l 阅读全文

posted @ 2020-07-04 16:58 cdekelon 阅读(444) 评论(0) 推荐(0) 编辑

SSD yolo
摘要:https://arxiv.org/abs/1708.02002 https://zhuanlan.zhihu.com/p/33544892 https://books.google.com.hk/books?id=OJXFDwAAQBAJ&pg=PA234&lpg=PA234&dq=coco+mo 阅读全文

posted @ 2020-06-23 16:45 cdekelon 阅读(57) 评论(0) 推荐(0) 编辑

dataset
摘要:http://groups.csail.mit.edu/vision/datasets/ADE20K/ https://pathmind.com/wiki/open-datasets https://www.kaggle.com/datasets/imneonizer/widerperson-kit 阅读全文

posted @ 2020-06-23 10:36 cdekelon 阅读(81) 评论(0) 推荐(0) 编辑

how to generate image labels after crop
摘要:https://blog.roboflow.ai/why-and-how-to-implement-random-crop-data-augmentation/ 阅读全文

posted @ 2020-06-09 14:54 cdekelon 阅读(41) 评论(0) 推荐(0) 编辑

1 2 3 下一页

导航

点击右上角即可分享
微信分享提示