ggml : remove inline specifier from fp16 <-> fp32 converters

diff --git a/ggml.c b/ggml.c
index c28628aa1c5b3f73dff9e15e36474eb706ceb179..b6d528d9fd6dd414351808cb648c5a774b95c70e 100644 (file)
--- a/ggml.c
+++ b/ggml.c
@@ -147,7 +147,7 @@ static inline uint32_t fp32_to_bits(float f) {
        return fp32.as_bits;
 }
 
-inline float ggml_fp16_to_fp32(ggml_fp16_t h) {
+float ggml_fp16_to_fp32(ggml_fp16_t h) {
     const uint32_t w = (uint32_t) h << 16;
     const uint32_t sign = w & UINT32_C(0x80000000);
     const uint32_t two_w = w + w;
@@ -170,7 +170,7 @@ inline float ggml_fp16_to_fp32(ggml_fp16_t h) {
     return fp32_from_bits(result);
 }
 
-inline ggml_fp16_t ggml_fp32_to_fp16(float f) {
+ggml_fp16_t ggml_fp32_to_fp16(float f) {
 #if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) || defined(__GNUC__) && !defined(__STRICT_ANSI__)
     const float scale_to_inf = 0x1.0p+112f;
     const float scale_to_zero = 0x1.0p-110f;