Efficient Methods for Natural Language Processing

1

Introduction

00:00 • 5min

2

How to Scale Out Efficiency in Deep Learning?

05:07 • 2min

3

Using 100% of the Data, You're Not Losing That Much

06:53 • 2min

4

Pre-Trained Models - Can I Apply This to Any Kind of Domain?

08:24 • 2min

5

The High Level Takeaway for Efficiency Methods

10:41 • 3min

6

Is There a Trade Off Between Expert and Mixed Approach?

13:49 • 2min

7

How Do You Reduce the Efficiency of Your Model?

15:43 • 2min

8

Pre-Training and Fine Tuning in NLP

17:14 • 2min

9

Using Labeled Examples to Fine Tune Your Model

19:10 • 3min

10

The Challenge of Fine Tuning GPUs

21:49 • 2min

11

Evaluating Models Like GPD-3

23:24 • 2min

12

NLP Models - What Are the First Two to Three Things You Would Try?

25:09 • 2min

13

NLP

27:04 • 1min

14

NLP, Data Set Creators and Data Set Engineers

28:34 • 2min

15

80% of Energy Costs Are on the Inference Side, Right?

30:42 • 2min

16

How Much Interest and Awareness Is There in Green AI?

32:57 • 2min

17

How Much It Costs to Train Large Neural Networks?

35:08 • 2min

18

Detoxification of Language Models

36:48 • 2min

19

Is There a Bias in Data Sets?

38:56 • 2min

20

Is There a Tool That Can Surface Biases?

40:53 • 2min

21

The Data Exchange Podcast Is a Property of Gradient Flow

43:12 • 2min