🚀 NanoHammer-1.5B-Instruct:

Show and Tell
1

We are excited to introduce NanoHammer, a novel architecture by NoesisLab designed for Causal State Compression and true Linear Inference Complexity.

:brain: The Core: Holographic State Space

Forget the growing KV Cache. NanoHammer leverages Holographic Rotary Embeddings to compress sequence history into a dynamic integral state.

  • Polynomial Compression: Instead of storing raw history, we “integrate” context into a complex number space (S_t = \\sum x_i e^{i\\theta}), treating memory as a container of evolving polynomial coefficients.

  • Dynamic Evolution: The architecture features a custom StateUpdateCell that uses Euler method fixed-point iteration, allowing the model to perform implicit reasoning via differential state updates.

:high_voltage: Why It Matters: Efficiency Meets Reasoning

  • O(1) Inference Memory: State size remains constant regardless of sequence length.

  • Causal Modeling: Explicitly models the causal flow of logic through time, perfect for “implicit reasoning” tasks without the verbosity of Chain-of-Thought.

  • 1.5B Lightweight Design: High performance, low resource footprint.

:hammer_and_wrench: Model Card Highlights

  • Type: nanohammer (Hybrid Causal-State Architecture)

  • License: Apache 2.0

  • Capabilities: Instruction following, Long-context handling

:link: Try it on Hugging Face: https://huggingface.co/NoesisLab/NanoHammer-1.5B-Instruct

1 Like