03 2019 档案
摘要:1,目录和前言 https://blog.csdn.net/itplus/article/details/37969519 2,预备知识:逻辑回归、贝叶斯公式、霍夫曼树 https://blog.csdn.net/itplus/article/details/37969635 3,背景知识:统计语言
阅读全文
摘要:https://realpython.com/python-speech-recognition/ The Ultimate Guide To Speech Recognition With Python by David Amos advanced data-science machine-lea
阅读全文
摘要:1,语音识别程序 https://blog.csdn.net/a18852867035/article/details/78565293 2,python语音识别指南终极版 https://blog.csdn.net/j2IaYU7Y/article/details/79878310 3,pytho
阅读全文
摘要:lane detection Paper 2019 2018 2017 Code Blog Dateset Paper 2019 《Robust Lane Detection from Continuous Driving Scenes Using Deep Neural Networks》 《En
阅读全文
摘要:0001, object-detection [TOC] This is a list of awesome articles about object detection. If you want to read the paper according to time, you can refer
阅读全文
摘要:A curated list of deep learning image classification papers and codes since 2014, Inspired by awesome-object-detection, deep_learning_object_detection
阅读全文
摘要:Semantic segmentation U-Net [https://arxiv.org/pdf/1505.04597.pdf] [2015] https://github.com/zhixuhao/unet [Keras] https://github.com/jocicmarko/ultra
阅读全文
摘要:【深度学习】目标检测算法总结(R-CNN、Fast R-CNN、Faster R-CNN、FPN、YOLO、SSD、RetinaNet) 目标检测是很多计算机视觉任务的基础,不论我们需要实现图像与文字的交互还是需要识别精细类别,它都提供了可靠的信息。本文对目标检测进行了整体回顾,第一部分从RCNN开
阅读全文
摘要:Object Detection with 10 lines of code Moses Olafenwa Jun 16, 2018 Moses Olafenwa Jun 16, 2018 Moses Olafenwa Jun 16, 2018 One of the important fields
阅读全文
摘要:基本知识 BOW,word2vector,glove cbow,skip-gram Hierarchical Softmax,Negative Sampling https://www.cnblogs.com/wkang/p/9611257.html https://www.cnblogs.com/
阅读全文
摘要:0000,标注工具 https://blog.csdn.net/chaipp0607/article/details/79036312 0001,概述 太优秀了,收藏用!转载自:https://www.cnblogs.com/Jie-Liang/archive/2017/06/29/6902375.
阅读全文
摘要:https://www.cnblogs.com/Determined22/p/6910277.html https://www.cnblogs.com/Determined22/p/6914926.html 原作者:Determined22
阅读全文
摘要:转自:https://blog.csdn.net/Chunfengyanyulove/article/details/86414810 简要概述文章精华 本篇文章主要解决了在目标检测中,检测框不是特别准,容易出现噪声干扰的问题,即close false positive,为什么会有这个问题呢?作者实
阅读全文
摘要:A,https://www.cnblogs.com/zhengzhe/p/7783270.html RCNN选择性搜索(Selective Search) RCNN选择性搜索(Selective Search) 基于: 1)图片大小 2)颜色 3)纹理 4)附件 算法一:分组分类算法 输入:(图层颜
阅读全文
摘要:https://blog.csdn.net/weixin_42398658/article/details/84639391 https://blog.csdn.net/qq_31050167/article/details/79161077 https://pan.baidu.com/s/1VQa
阅读全文
摘要:0001,常识1 计算机视觉的任务很多,有图像分类、目标检测、语义分割、实例分割和全景分割等,那它们的区别是什么呢?1、Image Classification(图像分类)图像分类(下图左)就是对图像判断出所属的分类,比如在学习分类中数据集有人(person)、羊(sheep)、狗(dog)和猫(c
阅读全文
摘要:1,原文:https://blog.csdn.net/u010725283/article/details/78593410 感受野(receptive field)被称作是CNN中最重要的概念之一。为什么要研究感受野呐?主要是因为在学习SSD,Faster RCNN框架时,其中prior box和
阅读全文
摘要:1,introduction Estimator 会封装下列操作: 训练 评估 预测 导出以供使用 预创建的 Estimator,也可以编写自定义 Estimator。所有 Estimator(无论是预创建的还是自定义)都是基于 tf.estimator.Estimator 类的类 2,Estima
阅读全文
摘要:1,tf-data两个新的抽象类 dataset表示一系列元素,其中每个元素包含一个或多个 Tensor 对象 创建来源(例如 Dataset.from_tensor_slices()),以通过一个或多个 tf.Tensor 对象构建数据集。 应用转换(例如 Dataset.batch()),以通过
阅读全文
摘要:1,Goodness An intuitive interface—Structure your code naturally and use Python data structures. Quickly iterate on small models and small data. Easier
阅读全文
摘要:1,Sequential model model = tf.keras.Sequential() # Adds a densely-connected layer with 64 units to the model:model.add(layers.Dense(64, activation='re
阅读全文
摘要:1,获取数据 imdb = keras.datasets.imdb(train_data, train_labels), (test_data, test_labels) = imdb.load_data(num_words=10000) 2,查看处理变形数据 2.1,查看 print(train_
阅读全文
摘要:1,数据集下载 2,一系列数据检查 这一条特别 3,Create feature columns and input functions,特征列和输入函数 3.1,one-hot-encoding, normalization, and bucketization 3.2,数字型和分类型 fc =
阅读全文
摘要:1,机器学习的基本步骤 Import and parse the data sets. Select the type of model. Train the model. Evaluate the model's effectiveness. Use the trained model to ma
阅读全文
摘要:1,一般描述 we saw that the accuracy of our model on the validation data would peak after training for a number of epochs, and would then start decreasing.
阅读全文
摘要:1,tf.layers基础函数 conv2d(). Constructs a two-dimensional convolutional layer. Takes number of filters, filter kernel size, padding, and activation funct
阅读全文
摘要:1,以类的方式定义一个模型 class Model(object): def __init__(self): # Initialize variable to (5.0, 0.0) # In practice, these should be initialized to random values
阅读全文
摘要:1,share的内容 code to create the model, and the trained weights, or parameters, for the model 2,ways There are different ways to save TensorFlow models—d
阅读全文
摘要:1,dataset的方法 Dataset.make_one_shot_iterator() or get_next() 2,使用python的方法-当eager mode enabled时 print('Elements of ds_tensors:')for x in ds_tensors: pr
阅读全文
摘要:map, batch, shuffle
阅读全文
摘要:1,几种方法 Create a source dataset using one of the factory functions like Dataset.from_tensors, Dataset.from_tensor_slices or using objects that read fro
阅读全文
摘要:1,tensor的特点 Tensors can be backed by accelerator memory (like GPU, TPU). Tensors are immutable 2,双向转换 TensorFlow operations automatically convert NumP
阅读全文
摘要:1, def get_flat_weights(model): weight_names = [ name for name in model.get_variable_names() if "linear_model" in name and "Ftrl" not in name] weight_
阅读全文
摘要:1, model_l1 = tf.estimator.LinearClassifier( feature_columns=base_columns + crossed_columns, optimizer=tf.train.FtrlOptimizer( learning_rate=0.1, l1_r
阅读全文
摘要:1,数字类型的 education_num = tf.feature_column.numeric_column('education_num')capital_gain = tf.feature_column.numeric_column('capital_gain')capital_loss =
阅读全文
摘要:1,简单数pandas import pandas train_df = pandas.read_csv(train_file, header = None, names = census_dataset._CSV_COLUMNS)test_df = pandas.read_csv(test_fil
阅读全文
摘要:1,25张图片 plt.figure(figsize=(10,10)) for i in range(25): plt.subplot(5,5,i+1) //plt.subplot(a,b,i+1) a*b>=25 plt.xticks([]) plt.yticks([]) plt.grid(Fal
阅读全文
摘要:https://tensorflow.google.cn/tutorials/representation/word2vec 暂时我们使用 vanilla 定义,将“上下文”定义为目标字词左侧和右侧的字词窗口 通过噪声对比训练进行扩展 神经概率语言模型一直以来都使用最大似然率 (ML) 原则进行训练
阅读全文