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 1 Packages Projects Releases Wiki Activity

dtype is copy assignable (#1436)

Browse Source
This commit is contained in:
Awni Hannun
2024-09-25 12:07:13 -07:00
committed by GitHub
parent 195b429d99
commit afc9c0ec1b
7 changed files with 29 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
mlx/io/gguf_quants.cpp
View File

@@ -120,7 +120,7 @@ void gguf_load_quantized(
std::vector<int> weights_shape = shape;
weights_shape.back() /= (weights_per_byte * 4);
auto w_nbytes = uint32.size *
auto w_nbytes = uint32.size() *
std::accumulate(weights_shape.begin(),
weights_shape.end(),
1,
@@ -130,7 +130,7 @@ void gguf_load_quantized(
// For scales and bias
shape[shape.size() - 1] = shape[shape.size() - 1] / weights_per_block;
auto sb_nbytes = float16.size *
auto sb_nbytes = float16.size() *
std::accumulate(shape.begin(), shape.end(), 1, std::multiplies<size_t>());
array scales(allocator::malloc(sb_nbytes), shape, float16);