Instructions to use state-spaces/mamba2-130m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use state-spaces/mamba2-130m with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("state-spaces/mamba2-130m", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Android + OpenCL implementation β runs on-device on non-flagship phones (Adreno 6xx)
#4
by a8nova - opened
Hi! I wanted to share an Android + OpenCL implementation of Mamba2-130M, in case anyone wants to run it on a phone:
- Try it: Edgi on Google Play β runs fully on-device, no cloud.
- Open source: the app is built on top of the open-source adreno-llms inference engine β https://github.com/a8nova/adreno-llms β pure C++/OpenCL with hand-written kernels tuned for Adreno, including custom kernels for the SSD recurrence and gated RMSNorm.
It's tuned and tested on Adreno 6xx GPUs β the GPU class in mid-range and older Android phones (verified on a 2020 Motorola Razr / Adreno 620) β and should run on most arm64 Android phones with OpenCL, though the optimizations are Adreno-specific. On the Adreno 620 it decodes at ~24.3 tok/s in fp16 β among the fastest models in the collection.
Hope it's useful β happy to answer questions!