@Daniel 深度碎片 第一课笔记汇总

[EmbraceLife]

fast.ai v3 2019中文笔记 fastai forum版 github 版

1. Lesson 0 How to get GPU running
   如何启动你的第一个GPU
2. What else do you need to get started?
   你需要做些什么准备？
3. How to make the most out of these lesson videos and notebooks?
   如何最大化利用课程视频与notebook?
4. what will you become and key resources? #66
   你可以期望的学习成就高度与关键学习资源
5. why should we learn from Jeremy Howard? #67
   为什么我们要跟着Jeremy Howard学习？
6. How to make DL accessible to everyone to do useful things? #68
   fast.ai让人人都成为深度学习高手的策略方针
7. How much to invest and What I get out? #69
   需要投入多少精力，会有怎样的收获？
8. Prerequisites and False assumptions and claims on DL #71 学习深度学习的必要基础是什么，以及常见的误解与成见
9. What will you be able to do at the end of lesson 1 #72 完成第一课后你将能做些什么
10. What is fastai learning philosophy #73 fast.ai的教学哲学是什么
11. How to use Jupyter notebook as a pro? #74 如何像专业人士一样使用jupyter notebook
12. What are Jupyter magics? #75 什么是Jupyter magics
13. What are fastai lib and how to use it? #76 什么是fastai库以及如何使用
14. Academic vs Kaggle Datasets, CatsDogs vs Pets dataset #77 学术和Kaggle数据，CatsDogs 与 Pets数据的异同
15. How to download dataset with fastai #78 如何下载数据包
16. How to access image folders and check filenames inside #79 如何进入图片文件夹以及查看里面的文件
17. How to get the labels of dataset? #80 如何从文件名中获取标注
18. Why and how to pick the image size for DataBunch? #81 为什么以及如何选择图片大小规格
19. What is a DataBunch #82 什么是DataBunch
20. What does normalize do to DataBunch? #83 如何normalize DataBunch
21. What to do if size is not 224? #84 如果图片大小不是224会怎样
22. What does it mean to normalize images #85 normalize 图片意味着什么
23. Why 224 not 256 as power of 2 to be image size #86 为什么图片尺寸是224而不是256
24. How to check the real images and labels #87 如何查看图片和标注
25. How to build a CNN learner/model? #88 如何构建一个CNN模型
26. Why use a pretrained model (framework and parameters) for your CNN? in other words, What is transfer learning? #89 为什么要用训练好的模型的框架与参数
27. what is overfitting? why wouldn’t the model cheating? #90 什么是过拟合？为什么我们的模型很难过拟合？
28. How to train the model with the best technique? #91 如何用最优的技术来训练模型？
29. How to find out how good is your model? #92 如何了解模型的好坏？
30. How to get the most out of this course? #93 如何放大课程效果？
31. The popularity of fastai library and replacing Keras #94 fastai的业界口碑以及与kera的对比
32. What students achieved with fastai and this course? #95 学员能够用fast.ai所学完成的项目
33. Why use Resnet rather than Inception? #96 为什么选择ResNet 而非Inception作为模型框架和已训练的参数
34. How to save a trained model? #97 如何保存训练好的模型
35. how to plot top losses examples/images? #98 如何画出损失值最高的数据/图？
36. How to find out the most confused images of our model? #99 如何找出模型混淆度最高的图片？
37. How to improve our model? finetuning #100 如何用微调改进模型？
38. what is CNN actually learning and why previous full model training didn’t work #101 CNN模型在学习些什么？以及为什么直接用全模型学习效果不佳？
39. How to train the whole model in the right way? #102 训练全模型的正确方式是什么？
40. How to improve model with more layers? #103
    如何用更大的模型来改进效果
41. Different ways to put your data into DataBunch #104 生成DataBunch的不同方式有哪些
42. How to make the most out of documents? #105 如何最大化利用fastai文档
43. QA on fastai with multi-GPU, 3D data #106 关于fastai与多GPU和3D数据的问题
44. An interesting and inspiring project #107 一些有趣的项目