grisuno/HPU-Core: release/v0.0.1

I report the discovery of a neural network state that I call a Hamiltonian Topological Crystal. By subjecting a Hamiltonian Processing Unit (HPU-Core) to extreme regularization pressure ($\lambda \approx 7.5 \times 10^{34}$), I observe a transition from conventional stochastic training to an integrable system with topological protection. The resulting model achieves perfect validation accuracy (1.0000) while operating in a "Vacuum Core" where 99.996% of parameters collapse toward zero. This work extends my previous research on Strassen crystallization from discrete algorithmic structures to continuous Hamiltonian dynamics, demonstrating that neural networks can encode physical laws as geometric structure rather than as data approximations. What's Changed Revise README with application log details by @m3rcyl3ss in https://github.com/grisuno/HPU-Core/pull/1 New Contributors @m3rcyl3ss made their first contribution in https://github.com/grisuno/HPU-Core/pull/1 Full Changelog: https://github.com/grisuno/HPU-Core/commits/release/v0.0.1