ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

llama : reorganize source code + improve CMake (#8006)

Browse source 
* scripts : update sync [no ci]

* files : relocate [no ci]

* ci : disable kompute build [no ci]

* cmake : fixes [no ci]

* server : fix mingw build

ggml-ci

* cmake : minor [no ci]

* cmake : link math library [no ci]

* cmake : build normal ggml library (not object library) [no ci]

* cmake : fix kompute build

ggml-ci

* make,cmake : fix LLAMA_CUDA + replace GGML_CDEF_PRIVATE

ggml-ci

* move public backend headers to the public include directory (#8122)

* move public backend headers to the public include directory

* nix test

* spm : fix metal header

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* scripts : fix sync paths [no ci]

* scripts : sync ggml-blas.h [no ci]

---------

Co-authored-by: slaren <slarengh@gmail.com>
This commit is contained in:
Georgi Gerganov 2024-06-26 18:33:02 +03:00 • committed by GitHub
parent 8854044561
commit f3f65429c4
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
345 changed files with 2555 additions and 1937 deletions
Download patch file Download diff file Expand all files Collapse all files

44
vulkan-shaders/rope_head.comp
View file

@ -1,44 +0,0 @@
#include "types.comp"
#extension GL_EXT_shader_16bit_storage : require
layout(local_size_x = 1, local_size_y = 256, local_size_z = 1) in;
layout (binding = 0) readonly buffer X {A_TYPE data_a[];};
layout (binding = 1) readonly buffer Y {int data_pos[];};
layout (binding = 2) readonly buffer Z {float data_ff[];};
layout (binding = 3) writeonly buffer D {D_TYPE data_d[];};
layout (push_constant) uniform parameter {
uint ncols;
uint n_dims;
float freq_scale;
uint p_delta_rows;
float freq_base;
float ext_factor;
float attn_factor;
float corr_dims[2];
float theta_scale;
uint has_ff;
} p;
float rope_yarn_ramp(const float low, const float high, const uint i0) {
const float y = (i0 / 2 - low) / max(0.001f, high - low);
return 1.0f - min(1.0f, max(0.0f, y));
}
void rope_yarn(const float theta_extrap, const uint i0, out float cos_theta, out float sin_theta) {
float mscale = p.attn_factor;
// Get n-d rotational scaling corrected for extrapolation
float theta_interp = p.freq_scale * theta_extrap;
float theta = theta_interp;
if (p.ext_factor != 0.0f) {
float ramp_mix = rope_yarn_ramp(p.corr_dims[0], p.corr_dims[1], i0) * p.ext_factor;
theta = theta_interp * (1 - ramp_mix) + theta_extrap * ramp_mix;
// Get n-d magnitude scaling corrected for interpolation
mscale *= 1.0f + 0.1f * log(1.0f / p.freq_scale);
}
cos_theta = cos(theta) * mscale;
sin_theta = sin(theta) * mscale;
}