code wiki / (root) / nx_gguf_typecensus.nx

nx_gguf_typecensus.nx

buildroot/runtime/nx_gguf_typecensus.nx

1562 B45 linesdepth 6pulls 9 transitivereach 0 importersview sourcekind tooltopic gguf
docsdependenciesstructsconstsfunctions

about

nx_gguf_typecensus.nx -- tally the real model's tensors by ggml_type, so we know what quant the weights actually are (the memory-bound decode lever depends on it: Q8_0=1B/val, Q4_K=0.56B/val, F32=4B/val). ggml types: 0=F32 1=F16 2=Q4_0 3=Q4_1 8=Q8_0 12=Q4_K 14=Q6_K. expect_exit: 0

dependencies 5 imports · 0 importers

nx_syscalls.nx nx_tier.nx nx_gguf.nx nx_gguf_load.nx nx_fmt.nx nx_gguf_typecensus.nx

imports: nx_syscalls.nxnx_tier.nxnx_gguf.nxnx_gguf_load.nxnx_fmt.nx

imported by: nobody (leaf or entry point)

call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown

main sys_mmap sys_read_file sys_openat_rd sys_lseek sys_mmap ↻ sys_read sys_close fmt_puts fmt_puts_fd sys_write nx_gguf_parse nx_le_read_u32 nx_le_read_u64 nx_le_read_u32 ↻ _gguf_skip_value nx_le_read_u64 ↻ nx_le_read_u32 ↻ _gguf_skip_value ↻ sys_mmap ↻ nx_gguf_tensor_at nx_gguf_tensor_n_values fmt_putn fmt_putn_fd sys_write ↻

structs

none

consts

none

functions

13func main() -> i64
calls 7: sys_mmapsys_read_filefmt_putsnx_gguf_parsenx_gguf_tensor_atnx_gguf_tensor_n_values+1