Yuyao Huang

Results 10 comments of Yuyao Huang

I played around with some ideas and implemented a VERY simple version of this feature. Hope anyone would feel interested in it . Demo file can be found in [My...

https://github.com/agentcooper/react-pdf-highlighter

I have had the same question and have convinced myself that the differences do exist. There are two weight W_k and W_v. W_k is for affinity generation, which is then...

Can you provide more details: 1. a full traceback print 2. the version of cutex, python and your OS

@hhaAndroid I read the focal loss paper which says it should be naturally extended to multi-class version, and the zhihu link seems to provide a correct implementation, which is very...

@jshilong May I have your suggestions on this issue? I have read #4559 and it would be very nice to have opinions from someone familiar with these lines of code....

![image](https://user-images.githubusercontent.com/17985266/176334434-84e29bbe-80b5-46a2-8c10-cd0eee833c08.png) Which "stars" each "circle" would average from?

Check if you have installed older version of `causal-conv1d` separately.

In the paper [Better plain ViT baselines for ImageNet-1k](https://arxiv.org/pdf/2205.01580), it says: > Notably absent from this baseline are further architectural changes, regularizers such as dropout or stochastic depth [7], ......

- [ ] https://www.semanticscholar.org/paper/InsightDrive%3A-Insight-Scene-Representation-for-Song-Guo/3d44626fe73c2b5ee301c36b61cc25500ad4caf9