-
Notifications
You must be signed in to change notification settings - Fork 48
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Adds gpu support for stateless llama #216
Conversation
IanNod
commented
Dec 1, 2023
- Adds support to compile vmfb to different device backends including cpu/vulkan/rocm/cuda
- Adds flags for device, iree_target_triple (backend device specific info) and vulkan max allocation
- Minor fix to gen_external_params when not doing quantization
@@ -48,6 +48,16 @@ | |||
"--precision", type=str, default="fp16", help="dtype of model [f16, f32]" | |||
) | |||
|
|||
parser.add_argument("--device", type=str, default="cpu", help="cpu, cuda, vulkan") |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
NIT: add rocm on here
I know this is a custom model, but would be nice to have a small test/runner for this, since this code can get a bit more complex. :) |
We have a very basic one that was added here cd063df but agreed we do need something more extensive |
- Adds support to compile vmfb to different device backends including cpu/vulkan/rocm/cuda - Adds flags for device, iree_target_triple (backend device specific info) and vulkan max allocation - Minor fix to gen_external_params when not doing quantization