Instructions to use kernels-community/mamba-ssm with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Kernels
How to use kernels-community/mamba-ssm with Kernels:
# !pip install kernels from kernels import get_kernel kernel = get_kernel("kernels-community/mamba-ssm") - Notebooks
- Google Colab
- Kaggle
Rebuild required for CUDA 13, B300 support
#4
by grg - opened
I'm running into issues with training Nemotron-Super with Axolotl on B300.
[rank5]: File "/root/.cache/huggingface/hub/models--kernels-community--mamba-ssm/snapshots/768416abf0ad7262a23d83c6efef10b9a4786489/build/torch210-cxx11-cu130-x86_64-linux/ops/triton/ssd_combined.py", line 772, in forward
[rank5]: assert causal_conv1d_cuda is not None, "causal_conv1d_cuda is not available. Please install causal-conv1d."
So the current kernel is not compatible with CUDA 13 B300, but this is already fixed in https://github.com/state-spaces/mamba/pull/733.
Could you rebuild the wheels please