zhangyiss/mlx
1
0
Fork 0
mirror of https://github.com/ml-explore/mlx.git synced 2025-12-16 01:49:05 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

Custom logsumexp (#2028)

Browse Source 
* initial custom logsumexp

* more tests

* comments + fix
This commit is contained in:
Awni Hannun
2025-03-31 07:36:55 -07:00
committed by GitHub
parent ec2854b13a
commit de5f38fd48
27 changed files with 590 additions and 255 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
mlx/backend/metal/softmax.cpp
View File

@@ -23,12 +23,7 @@ void Softmax::eval_gpu(const std::vector<array>& inputs, array& out) {
// Make sure that the last dimension is contiguous
auto set_output = [&s, &out](const array& x) {
bool no_copy = x.flags().contiguous && x.strides()[x.ndim() - 1] == 1;
if (no_copy && x.ndim() > 1) {
auto s = x.strides()[x.ndim() - 2];
no_copy &= (s == 0 || s == x.shape().back());
}
if (no_copy) {
if (x.flags().contiguous && x.strides()[x.ndim() - 1] == 1) {
if (x.is_donatable()) {
out.copy_shared_buffer(x);
} else {