MCPcopy Create free account
hub / github.com/andrewkchan/deepseek.cpp / float_to_half

Function float_to_half

src/codec.h:26–28  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

24 return _cvtsh_ss(x);
25}
26inline f16_t float_to_half(float x) {
27 return _cvtss_sh(x, 0);
28}
29#else
30inline float half_to_float(f16_t x) {
31 assert(false && "float16 not supported on this platform");

Callers 7

quantize_row_q2_K_refFunction · 0.85
quantize_row_q3_K_refFunction · 0.85
ropeFunction · 0.85
rope_v3Function · 0.85
_attention_implMethod · 0.85
float_array_to_halfFunction · 0.85
float_to_float8e5m2Function · 0.85

Calls

no outgoing calls

Tested by 1

float_array_to_halfFunction · 0.68