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

ggml : fix YARN + add tests + add asserts (#7617)

Browse source 
* tests : add rope tests

ggml-ci

* ggml : fixes (hopefully)

ggml-ci

* tests : add non-cont tests

ggml-ci

* cuda : add asserts for rope/norm + fix DS2

ggml-ci

* ggml : assert contiguousness

* tests : reduce RoPE tests

ggml-ci
This commit is contained in:
Georgi Gerganov 2024-05-29 20:17:31 +03:00 • committed by GitHub
parent cce3dcffc5
commit fb76ec31a9
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
12 changed files with 167 additions and 104 deletions
Download patch file Download diff file Expand all files Collapse all files

4
ggml_vk_generate_shaders.py
View file

@ -2670,14 +2670,12 @@ void main() {
const uint i = row*p.ncols + ib*p.ndims + ic/2;
const uint i2 = row/p.p_delta_rows;
const float cur_rot = p.inv_ndims * ic - ib;
const int pos = data_b[i2];
const float freq_factor = p.has_freq_facs != 0 ? data_freq_factors[ic/2] : 1.0f;
const float theta_base = pos*p.freq_scale*pow(p.theta_scale, col/2.0f) / freq_factor;
float cos_theta, sin_theta;
rope_yarn(theta_base, uint(cur_rot), cos_theta, sin_theta);
rope_yarn(theta_base, ic, cos_theta, sin_theta);
const float x0 = float(data_a[i + 0]);
const float x1 = float(data_a[i + p.ndims/2]);