Vulkan Bugfixes and Improvements (#7084)

Browse Source

* Modify mat mat mul shader for mul_mat_id, modify mat vec mul shaders for single call batch operation

* Further work towards MoE, disabled for now

* Disable MoE code (not ready yet), fix a number of bugs in shaders and Vulkan code

* Add softmax with f16 mask and pos buffer support

* Disable mul_mat_id shaders for now

* Fix flake8

* Fix validation errors caused by empty buffers on larger batch sizes

...

This commit is contained in:

0cc4m 2024-05-09 20:39:54 +02:00 committed by GitHub

parent d46dbc76f8

commit befddd0f15

No known key found for this signature in database

GPG Key ID: B5690EEEBB952194

3 changed files with 43648 additions and 34961 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Show Stats Download Patch File Download Diff File Expand all files Collapse all files

76726

ggml-vulkan-shaders.hpp

View File

File diff suppressed because it is too large Load Diff

1167

ggml-vulkan.cpp

View File

File diff suppressed because it is too large Load Diff

716

ggml_vk_generate_shaders.py

View File

File diff suppressed because it is too large Load Diff