-
Thank you very much for your excellent work! I have a question regarding the article titled "Frequency-domain MLPs are More Effective Learners in Time Series Forecasting." Could you please explain the…
-
The current triplane-mlp implementation, without any regularization, has decent performance, but shows density floater artifacts. Attached qualitative on Hotdog.
Hotdog results:
```
Mean PSNR on …
-
How would you replicate concatenated/merged layers?
-
multi-layer perceptrons
-
I updated the plugin recently and have found that attempts to merge any LoRA now throw the following IndexError:
```
Traceback (most recent call last):
File "D:\Lab\sd\a1111-xl\extensions\sd-webu…
-
Referring to the code usage examples, I think the usage methods of fastkan and efficient_kan should be consistent. I tried to replace MLP with KAN in my model. Using efficient_kan is normal, but using…
-
![image](https://user-images.githubusercontent.com/8708551/178094404-7150aa5f-2de8-4e5c-8b92-d770998ecbbc.png)
-
Hi! I made a test of how KAN could assist denoising diffusion and adapted the very initial spiral-diffusion model utilizing an MLP to KAN.
You can see that a two layer KAN fares almost as good as a…
-
I want to share this repo where we evaluate ACC performance between KAN and MLP in four scenarios: RAW Multiclass, RAW Binary, CSP Multiclass, and CSP Binary.
Draw your conclusions (see ours in the…
-
大佬你们好,请问使用48GB的单张A40显卡能否在scannet数据集上训练PTv3(没有使用flash attention),初始的batchsize为12时,发生了爆显存的情况,改成6依旧爆显存,将其改为1才能开始训练。
下面是详细的报错情况
(pointcept) root@1sq9f61k2p5u-0:~/zwj/Pointcept-main# sh scripts/train.sh …