BrainAI Multi-Modal Encoder Suite

What problem does it solve?

BrainAI's encoder suite uses heterogeneous per-modality outputs that must be standardized to a common workspace for reliable downstream reasoning. This skill provides adapters, a contract-driven validation framework, and tooling to enforce a uniform (B, T, D) EncoderOutput across vision, text, audio, sensors, and engram encoders.

Core Features & Use Cases

• Unified EncoderOutput contract: every modality outputs (B, T, D) with a shared mask and optional fields, enabling deterministic concatenation and modular workloads.
• Adapter framework and registry: per-modality adapters that project internal encoder outputs to the shared workspace, preserving timing and optional metadata.
• Dependency-fallback and testing suite: automated tests and tooling to validate shapes, dtypes, variable lengths, and resilience to missing dependencies.

Quick Start

Run the encoder-suite tests to validate all modality adapters against the EncoderOutput contract.