Why my controlnet output = 0?

Nobody has claimed this yet.

Hi, I am training a new network based on Controlnet with FFHQ dataset. But when i was checking my code, I find that the output of 'guided_hint = self.input_hint_block(hint, emb, context)' equals 0. I will be grateful if you can give me some advice.

The result is from:

self.input_hint_block = TimestepEmbedSequential(
conv_nd(dims, hint_channels, 16, 3, padding=1),
nn.SiLU(),
conv_nd(dims, 16, 16, 3, padding=1),
nn.SiLU(),
conv_nd(dims, 16, 32, 3, padding=1, stride=2),
nn.SiLU(),
conv_nd(dims, 32, 32, 3, padding=1),
nn.SiLU(),
conv_nd(dims, 32, 96, 3, padding=1, stride=2),
nn.SiLU(),
conv_nd(dims, 96, 96, 3, padding=1),
nn.SiLU(),
conv_nd(dims, 96, 256, 3, padding=1, stride=1),
nn.SiLU(),
zero_module(conv_nd(dims, 256, model_channels, 3, padding=1))
)

Start by tracing the input_hint_block definition and the zero_module call shown in the issue, then inspect how guided_hint is produced during ControlNet training. Compare the tensor values before and after this block and determine whether the zero output is expected initialization or a defect; done means identifying the cause and documenting the required change or usage.

Written by the indexing model from the issue text.