Qianli Ma
Results
13
comments of
Qianli Ma
I got the same error, thanks to your solution which resolve it! > I’m using torch==2.2.2 which fixed the problem.
> 兄弟们，只要把modelng_t5.py的264-275行注释掉就infer起来了 我之前尝试把FusedRMSNorm替换成nn.LayerNorm，结果sample的输出全是nan🤡
ReDDiT: Efficient Diffusion as Low Light Enhancer Paper: https://arxiv.org/abs/2410.12346 Code: https://github.com/lgz-0713/ReDDiT Topic: Low-level vision, Image enhancement