nx_t2mesh.nx
buildroot/runtime/nx_t2mesh.nx
about
nx_t2mesh.nx -- sovereign TEXT-to-MESH generator v0 (generator-path R2). NOT a neural DreamFusion/Shap-E
model -- a DATA-DRIVEN PARAMETRIC generator: a prompt selects an object TEMPLATE (a recipe of SDF primitives),
which fills the field; nx_meshgen (surface nets) then EMITS the mesh. The point vs hand-crafting: ONE call
emits N different objects from N prompts, and the vocabulary GROWS AS DATA (add a template = add an object).
The critic scores each emitted object. Honest ceiling: parametric (not free-form), a spine to grow toward
neural later. license_tier: ORIGINAL
dependencies 2 imports · 4 importers
imports: nx_syscalls.nxnx_sdfrender.nx
imported by: nx_autorig_gate.nxnx_gen_orchestrator.nxnx_genloop_gate.nxnx_t2mesh_gate.nx
structs
| none |
consts
| 9 | const K_MAGIC_1050: i64 = 1050 |
functions
| 11 | func t2m_set(base: i64, idx: i64, cx: i64, cy: i64, cz: i64, rx: i64, ry: i64, rz: i64) -> i64 called by 1: t2m_gen |
| 17 | func t2m_conf(base: i64, count: i64, blend: i64) -> i64 |
| 25 | func t2m_has(s: *u8, kw: *u8) -> i64 called by 1: t2m_id_of |
| 42 | func t2m_id_of(prompt: *u8) -> i64 |
| 60 | func t2m_name(id: i64) -> *u8 |
| 73 | func t2m_gen(base: i64, id: i64) -> i64 |
| 123 | func t2m_prompt(base: i64, prompt: *u8) -> i64 { return t2m_gen(base, t2m_id_of(prompt)) } |