Mamba
Collection
Mamba is a new LLM architecture that integrates the Structured State Space sequence model to manage lengthy data sequences. • 8 items • Updated • 2
The Model mlx-community/mamba-370m-hf-f32 was converted to MLX format from state-spaces/mamba-370m-hf using mlx-lm version 0.18.1.
pip install mlx-lm
from mlx_lm import load, generate
model, tokenizer = load("mlx-community/mamba-370m-hf-f32")
response = generate(model, tokenizer, prompt="hello", verbose=True)
Quantized
Base model
state-spaces/mamba-370m-hf