-
Notifications
You must be signed in to change notification settings - Fork 936
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
triton error while running Mamba2 with slow path #369
Comments
Tried the following and time seems not to change. Maybe this is just an initial delay:
|
Well I’m wondering about why adding compile as #355 discussion makes the code failed, as the author mentioned this could accelerate a lot |
the same issue |
同样的问题 |
the same issue |
Hi, I have the same problem, have you solved it? |
the same issue |
as #355 , I added "@torch.compile(options={"triton.cudagraphs": True}, fullgraph=True)" to "mamba_chunk_scan_combined" function in file "ssd_combined.py", and running failed with error:
reproduce code:
I'm not sure what to provide, but my packages are:
mamba-ssm 2.0.3
causal-conv1d 1.2.2.post1
pytorch 2.3.1 with py39_cu121_cudnn8.9.2_0
The text was updated successfully, but these errors were encountered: