# Eddie's Learning Record 18

**Published by:** [Eddie's Learning Records](https://paragraph.com/@eddiehe/)
**Published on:** 2022-10-17
**URL:** https://paragraph.com/@eddiehe/eddies-learning-record-18

## Content

1. DurationMonday, September 26th, 2022 - Friday, September 30th, 2022 2. Learning Record2.1 Paper WorkI spent two days finishing the proposal for my project. It was really boring. But preparing the PowerPoint is definitely a good way to organize my thought. I finished the pitching on the afternoon of Wednesday, September 28th. Unexpectedly, it was peace & love. The teachers asked me questions with kindness instead of judging me. This activity warned me that I needed to hurry up as most of the other students had finished the majority of their experiments. However, I just finished the preparatory search. 2.2 Learning AttentionI learned how to deploy attention modules from bilibili. The tutorial teaches me to use a 6 G voice dataset, which shocked me. There is no need to download gigabytes of data just for learning attention. I truncated the data into an about 85 M mini dataset and it worked well. Not only did I have a brief concept of the attention mechanism, but also I found it cool to use tqmd in PyTorch. 3. Feeling3.1 So SoI have no special feelings this week. I only knew what I learned now was far away from the technique used in updated papers. Consequently, I lost my passion for diving into each code like I used to.

## Publication Information

- [Eddie's Learning Records](https://paragraph.com/@eddiehe/): Publication homepage
- [All Posts](https://paragraph.com/@eddiehe/): More posts from this publication
- [RSS Feed](https://api.paragraph.com/blogs/rss/@eddiehe): Subscribe to updates
- [Twitter](https://twitter.com/eddiehe99): Follow on Twitter

## Optional

- [Collect as NFT](https://paragraph.com/@eddiehe/eddies-learning-record-18): Support the author by collecting this post
- [View Collectors](https://paragraph.com/@eddiehe/eddies-learning-record-18/collectors): See who has collected this post