issues
search
alxndrTL
/
mamba.py
A simple and efficient Mamba implementation in pure PyTorch and MLX.
MIT License
960
stars
86
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
About the speed test
#8
MzeroMiko
closed
8 months ago
9
MLX: workaround for torch.BFloat16Tensor + get rid of Python generator in mamba_lm_mlx.MambaLM.generate for UTF-8 multibyte characters like CJK, Thai and etc.
#7
beebopkim
closed
8 months ago
2
MLX inference error with BFloat16
#6
beebopkim
closed
8 months ago
7
MLX memory usage at inference
#5
alxndrTL
opened
8 months ago
0
mlx support : pscan, mamba and mamba_lm files rewritten in MLX
#4
alxndrTL
closed
9 months ago
0
AR Inference - step function
#3
alxndrTL
closed
9 months ago
1
training functions?
#2
win10ogod
closed
8 months ago
4
Refactor mamba to integrate torch.compile, reference invocation, accuracy errors
#1
voznesenskym
opened
9 months ago
0
Previous