nx_nofloat_qwen_encode_gate.nx
buildroot/runtime/nx_nofloat_qwen_encode_gate.nx
about
nx_nofloat_qwen_encode_gate.nx -- BPE ENCODE: text -> token ids, sovereign, from the real GGUF merges+vocab.
Algorithm: start with single-char tokens; repeatedly merge the adjacent pair (X,Y) whose rule "X Y" has the
LOWEST rank in tokenizer.ggml.merges; stop when no adjacent pair has a merge; map each final token-string to
its id. (ASCII input -> byte-level pre-tokenize is identity, so we encode "hello" directly.) Proven by
ROUND-TRIP: the encoded tokens' strings concatenate back to the input, and every token maps to a valid id
(so decode(ids) == input). Pure metadata + byte ops (nx_gguf_meta). No hw writes (Rule 26).
expect_exit: 0 license_tier: ORIGINAL
dependencies 7 imports · 0 importers
imports: nx_syscalls.nxnx_tier.nxnx_le.nxnx_tensor.nxnx_gguf.nxnx_gguf_meta.nxnx_gate_verdict.nx
imported by: nobody (leaf or entry point)
call flow from main pre-order; caps 40 nodes / depth 6 declared; ↻ = already shown
structs
| none |
consts
| none |
functions
| 16 | func en_puts(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} sys_write(1,s,n); return 0 } |
| 17 | func en_num(v: i64) -> i64 { let b: *u8=sys_mmap(28); var m: i64=v; if m<0{m=0-m;sys_write(1,"-" as *u8,1)} let t: *u8=sys_mmap(28); var k: i64=0; if m==0{t[0]=48 as u8;k=1} while m>0{t[k]=(48+(m%10)) as u8;m=m/10;k=k+1} var i: i64=0; while i<k{b[i]=t[k-1-i];i=i+1} sys_write(1,b,k); return 0 } |
| 18 | func en_slen(s: *u8) -> i64 { var n: i64=0; while s[n]!=(0 as u8){n=n+1} return n } called by 1: main |
| 19 | func bytes_eq(a: *u8, b: *u8, n: i64) -> i64 { var i: i64=0; while i<n { if a[i]!=b[i] { return 0 } i=i+1 } return 1 } |
| 20 | func find_token_id(buf: *u8, first: i64, vocab: i64, target: *u8, tlen: i64) -> i64 |
| 26 | func merge_rank(buf: *u8, mfirst: i64, nm: i64, xp: *u8, xl: i64, yp: *u8, yl: i64, tmp: *u8) -> i64 |
| 37 | func main() -> i64 |