nx_text2motion.nx

nx_text2motion.nx -- TEXT-CONDITIONED neural motion (the MotionGPT headline, S6b): a description -> motion, via LEARNED weights, no keyword if-else. Two learned stages, sovereign + integer-deterministic: (1) TEXT ENCODER: bag-of-words over a small vocab -> a linear classifier (weight matrix W[class][word]) TRAINED by the perceptron rule (error-driven weight updates) on (caption -> move-class) pairs. This is real supervised learning; the honest test is GENERALIZATION to captions never trained on (held-out). (2) CLASS-CONDITIONED generator: per class, a token-transition model learned from that class's motion tokens (composes with nx_motion_neural's VQ codebook), so text -> class -> a generated motion-token sequence. The chain: caption -> t2m_encode (learned argmax) -> class -> t2m_gen_tokens (learned transitions) -> decode via the VQ codebook -> drive the rig. Fixed vocab + fixed classes = data; grow by adding captions/moves. license_tier: ORIGINAL