Spaces:
Build error
Build error
| // These tests may take a long time! | |
| // They are to prove that conversion from double to float of various functions in ggml.c doesn't affect the result. | |
| // This is done by checking all finite (non-NaN, non-infinite) floats. | |
| // ggml.c::quantize_row_q4_0_ref | |
| inline static uint8_t round_orig(float v0) { return ((int8_t) (round(v0))) + 8; } | |
| // ggml.c::ggml_silu_f32 | |
| inline static float silu_orig(float x) { | |
| return x/(1.0 + exp(-x)); | |
| } | |
| // ggml.c::quantize_row_q4_0_ref | |
| inline static uint8_t round_float(float v0) { return (int8_t)roundf(v0) + 8; } | |
| // ggml.c::ggml_silu_f32 | |
| inline static float silu_float(float x) { | |
| return x/(1.0f + expf(-x)); | |
| } | |
| int main(void) { | |
| uint32_t x = UINT32_MAX; | |
| do { | |
| float f; | |
| memcpy(&f, &x, sizeof(x)); | |
| assert(!std::isfinite(f) || (round_orig(f) == round_float(f))); | |
| } while (x--); | |
| // GELU and SILU implementations are used with a FP16 lookup table. | |
| // The original and float-only results are not equal for all inputs after converting to FP16. | |
| // GELU is an approximation anyway (tanh), not tested here. | |
| // For SILU, verify that the results are at least the closest floating point numbers, if the FP16 values don't match. | |
| for (x = 0; x <= UINT16_MAX; x++) { | |
| float f = _cvtsh_ss(x); | |
| const float so = silu_orig(f); | |
| const float sf = silu_float(f); | |
| assert( (_cvtss_sh(so, 0) == _cvtss_sh(sf, 0)) | |
| || (nextafterf(so, sf) == sf) | |
| || (nextafterf(sf, so) == so)); | |
| } | |
| } | |