filmov
tv
Lesson 4: Practical Deep Learning for Coders
Показать описание
COLLABORATIVE FILTERING, EMBEDDINGS, AND MORE
When we ran this class at the Data Institute, we asked what students were having the most trouble understanding, and one of the answers that kept coming up was “convolutions”. So, we start with a detailed look at the convolution operation, by implementing a couple of convolutional layers and filters in a spreadsheet. Next up, we give SGD (and modern accelerated variants) the same treatment. Once you’ve seen how easy accelerated SGD methods are, try reading the original papers—notice how even the most complex deep learning papers tend to look simple once you’ve digested and implemented them?
Then we look further into avoiding over-fitting, and learn a clever trick for datasets where you have a lot more unlabeled data than labeled data (that is, semi-supervized learning scenarios): “pseudo-labeling” and “knowledge distillation”.
Finally, we move away from computer vision for the first time, to discussion recommendation systems, and in particular, collaborative filtering techniques. This is both a useful technique of itself, and will also be a great introduction to embeddings, which is going to be critical when we learn about natural language processing in the next lesson.
When we ran this class at the Data Institute, we asked what students were having the most trouble understanding, and one of the answers that kept coming up was “convolutions”. So, we start with a detailed look at the convolution operation, by implementing a couple of convolutional layers and filters in a spreadsheet. Next up, we give SGD (and modern accelerated variants) the same treatment. Once you’ve seen how easy accelerated SGD methods are, try reading the original papers—notice how even the most complex deep learning papers tend to look simple once you’ve digested and implemented them?
Then we look further into avoiding over-fitting, and learn a clever trick for datasets where you have a lot more unlabeled data than labeled data (that is, semi-supervized learning scenarios): “pseudo-labeling” and “knowledge distillation”.
Finally, we move away from computer vision for the first time, to discussion recommendation systems, and in particular, collaborative filtering techniques. This is both a useful technique of itself, and will also be a great introduction to embeddings, which is going to be critical when we learn about natural language processing in the next lesson.
1:34:37
Lesson 4: Practical Deep Learning for Coders 2022
2:03:47
Lesson 4: Practical Deep Learning for Coders
2:14:32
Lesson 4: Deep Learning 2018
1:42:49
Lesson 5: Practical Deep Learning for Coders 2022
2:01:35
Lesson 4 - Deep Learning for Coders (2020)
2:51:40
Practical Deep Learning for Coders - Lesson 4 - Part 1
1:43:39
Lesson 4: Deep Learning 2019 - NLP; Tabular data; Collaborative filtering; Embeddings
1:20:51
Practical Deep Learning for Coders - Lesson 4 - Part 2
0:18:27
DO NOT MAKE CHEAP OFFERING. | HOW NOT TO GIVE | LESSON 4
1:30:25
Lesson 3: Practical Deep Learning for Coders 2022
11:12:32
Practical Deep Learning for Coders - Full Course from fast.ai and Jeremy Howard
2:03:18
Lesson 3: Practical Deep Learning for Coders
0:55:31
Lesson '0': Practical Deep Learning for Coders (fast.ai)
2:15:16
Lesson 9: Deep Learning Foundations to Stable Diffusion
1:42:55
Lesson 6: Practical Deep Learning for Coders 2022
0:10:00
#16 Machine Learning Specialization [Course 1, Week 1, Lesson 4]
1:36:55
Lesson 8 - Practical Deep Learning for Coders 2022
1:22:56
Practical Deep Learning for Coders: Lesson 1
1:16:42
Lesson 2: Practical Deep Learning for Coders 2022
0:05:52
Deep Learning | What is Deep Learning? | Deep Learning Tutorial For Beginners | 2023 | Simplilearn
2:03:04
Lesson 5: Practical Deep Learning for Coders
0:00:16
How to eat Roti #SSB #SSB Preparation #Defence #Army #Best Defence Academy #OLQ
0:04:42
04: Lesson-4 NLP:Tabular Data; Recsys | fast.ai 2019 & Things Jeremy Howard says to do
2:14:31
Intro to Deep Learning 2018 - Lesson 4
Комментарии