Skip to content

Conversation

@lukicdarkoo
Copy link
Member

@lukicdarkoo lukicdarkoo commented Aug 16, 2025

Copilot suggestions:

  • Switch to attention-based multi-view aggregation by default
  • Strengthen modality-specific heads (image, depth, vector) with GELU, LayerNorm, and optional dropout
  • Improve vector input handling (batched and list forms)
  • Normalize fused features before the final head
  • Add careful initialization for newly added layers

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

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants