Set the initial RGB processing module based on the backbone type.

Verify the validity of the input arguments to ensure consistent behaviour. Specifically when selecting neurons for sychronisation us

Verify the validity of the input arguments.

Args: index (int): Index Returns: tuple: (sample, target) where target is class_index of the target class.

Args: optimizer (torch.optim.Optimizer): 最適化手法インスタンス warmup_epochs (int): linear warm

Returns an iterator that yields 'num_samples_per_epoch' indices. It uses itertools.islice to take a finite slice from the inf

The number of samples produced by the iterator per __iter__ call.

Override to handle lazy weights initialization.

Forward pass - Reverted to structure closer to user's working version. Executes T=iterations steps.

Forward pass - Reverted to structure closer to user's working version. Executes T=iterations steps.

Args: x (torch.Tensor): Input tensor, expected shape (B, N, in_dims) where B=batch, N=d_model, in_d

Maps -1 (negative parity) to 0 and 1 (positive) to 1

Computes positional encodings for the input feature map x. Args: x (torch.Tensor): Input feature map, shape (B, C, H, W)

Computes combined positional encoding. Args: x (torch.Tensor): Input feature map, shape (B, C, H, W). Returns:

Computes rotational positional embeddings based on input width. Args: x (torch.Tensor): Input tensor (used for shape and

Forward pass - Reverted to structure closer to user's working version. Executes T=iterations steps.

Get the dimensionality of the backbone output.

Get the dimensionality of the backbone output.

Get the positional embedding module.

Get the synapse module. We found in our early experimentation that a single Linear, GLU and LayerNorm block performed worse than two