From 27b69c2b75a251ec39ece318c26ccb1a25d9fd79 Mon Sep 17 00:00:00 2001 From: YurenHao0426 Date: Sat, 23 May 2026 04:37:43 -0500 Subject: Add SRM-Joint-AOL v1 model MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Forked from HRM ACT framework. Replaces dual H_level/L_level transformer stacks with a single joint operator T(h, l) that is provably contractive under weighted P-norm with Lip_P(T) ≤ (1-α) + α·κ < 1. Per-step Lyapunov bound: λ_1 ≤ log((1-α) + α·κ). With κ=0.86 → λ_1 ≤ -0.15 ≈ HRM success regime (no CF needed). Components: - AOLLinear: 1-Lipschitz via Prach & Lampert rescaling (float32 normalization) - AOLBlock: stack with ReLU (1-Lipschitz activation) - CayleyOrthogonal: exact orthogonal cross-coupling - BlockGain: softmax row-sum bound under P-norm - AOLTokenMixer: 1-Lipschitz token + channel mixing Smoke test passes: params=939k (hidden=256 test config), forward OK, empirical Lip=0.17 < theoretical bound 0.90. Co-Authored-By: Claude Opus 4.7 --- models/srm/__init__.py | 0 1 file changed, 0 insertions(+), 0 deletions(-) create mode 100644 models/srm/__init__.py (limited to 'models/srm/__init__.py') diff --git a/models/srm/__init__.py b/models/srm/__init__.py new file mode 100644 index 0000000..e69de29 -- cgit v1.2.3