metal : alibi for arbitrary number of heads (#3426)

This commit is contained in:
Jiahao Li 2023-10-04 00:55:21 +08:00 committed by GitHub
parent 017efe899d
commit f56e1baec3
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 13 additions and 7 deletions

View file

@ -830,7 +830,9 @@ kernel void kernel_alibi_f32(
constant uint64_t & nb1,
constant uint64_t & nb2,
constant uint64_t & nb3,
constant float & m0,
constant float & m0,
constant float & m1,
constant int & n_heads_log2_floor,
uint3 tgpig[[threadgroup_position_in_grid]],
uint3 tpitg[[thread_position_in_threadgroup]],
uint3 ntg[[threads_per_threadgroup]]) {
@ -846,7 +848,12 @@ kernel void kernel_alibi_f32(
const int64_t i0 = (n - i3*ne2*ne1*ne0 - i2*ne1*ne0 - i1*ne0);
device float * dst_data = (device float *) ((device char *) dst + i3*nb3 + i2*nb2 + i1*nb1 + i0*nb0);
float m_k = pow(m0, i2 + 1);
float m_k;
if (i2 < n_heads_log2_floor) {
m_k = pow(m0, i2 + 1);
} else {
m_k = pow(m1, 2 * (i2 - n_heads_log2_floor) + 1);
}
for (int64_t i00 = tpitg.x; i00 < ne00; i00 += ntg.x) {
device const float * src = (device float *)((device char *) src0 + i03*nb03 + i02*nb02 + i01*nb01 + i00*nb00);
dst_data[i00] = src[0] + m_k * (i00 - ne00 + 1);