Skip to content

Question about _compress_ar in JointAutoregressiveHierarchicalPriors #332

Description

@chunbaobao

Sorry to bother you, but I have some stupid questions. I noticed that the _compress_ar function in the JointAutoregressiveHierarchicalPriors model uses two for-loops during the encoding process to predict the distribution of y_hat for all individual pixels:

masked_weight = self.context_prediction.weight * self.context_prediction.mask
for h in range(height):
    for w in range(width):
        y_crop = y_hat[:, :, h : h + kernel_size, w : w + kernel_size]
        ctx_p = F.conv2d(
            y_crop,
            masked_weight,
            bias=self.context_prediction.bias,
        )

I was wondering if this could be implemented in parallel, similar to how the training process get the scales and means for all pixels simultaneously? (Decoding must be implemented serially.)

y = self.g_a(x)
z = self.h_a(y)
z_hat, z_likelihoods = self.entropy_bottleneck(z)
gaussian_params = self.h_s(z_hat)
scales_hat, means_hat = gaussian_params.chunk(2, 1)

Would implementing this in parallel cause any issues, such as inconsistency between training and runtime?

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions