vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Fix documentation

Browse source
This commit is contained in:
Justine Tunney 2024-04-21 14:18:00 -07:00
parent ce0442d7a7
commit bc278c8a0e
No known key found for this signature in database
GPG key ID: 52965314629936D4
1 changed files with 3 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

13
ggml-impl.h
View file

@ -18,7 +18,9 @@
#define MAX(a, b) ((a) > (b) ? (a) : (b)) #define MAX(a, b) ((a) > (b) ? (a) : (b))
/** /**
* Google Brain 16-bit floating point number. * Converts brain16 to float32.
*
* The bfloat16 floating point format has the following structure:
* *
* ┌sign * ┌sign
* │ * │
@ -52,17 +54,8 @@
* │┌─┴─┐┌─┴──────┐ * │┌─┴─┐┌─┴──────┐
* 0b0000000000000000 IEEE binary16 * 0b0000000000000000 IEEE binary16
* *
* So be warned that converting between them, destroys several bits.
*
* @see IEEE 754-2008 * @see IEEE 754-2008
*/ */
struct ggml_bf16_s {
uint16_t bits;
};
/**
* Converts brain16 to float32.
*/
static inline float ggml_compute_bf16_to_fp32(ggml_bf16_t h) { static inline float ggml_compute_bf16_to_fp32(ggml_bf16_t h) {
union { union {
float f; float f;