Li Jun
Li Jun
> > It uses the Gumble-softmax to make propagation come true. > > May I ask on which line of the code is it I am also confused about this.
@anxiangsir 请问pretrain详细的README在哪呢？
是这样的，测试时候用整张图才是合理的
如果是对整个图像做的那么大图的hw与训练时候的hw不一致，也就是说有一个Mamba块的参数发生了变化 ```python self.model3 = Mamba( # This module uses roughly 3 * expand * d_model^2 parameters d_model=w*h, # Model dimension d_model d_state=16, # SSM state expansion factor d_conv=4, # Local...
请问作者是如何处理的呢？
@zzr-idam 刚刚又看了代码发现是针对一个patch在测试，这严谨吗？
Sincere thanks for your prompt response. In my opinion, `torch.compile` is more convenient and faster compared to `torch.jit`. I believe that once `ptwt` supports `torch.compile`, it will be even more...
Torch.compile() indicates that some operations in ptwt are not supported. ---Original--- From: "Moritz ***@***.***> Date: Thu, Aug 22, 2024 19:54 PM To: ***@***.***>; Cc: ***@***.******@***.***>; Subject: Re: [v0lta/PyTorch-Wavelet-Toolbox] Support torch.compile()?...
Could you provide an example of how to apply learnable wavelet transforms to a 3D tensor? The input has the shape [B, L, C], where B: batch_size, L: length, and...