so i repaired Zonos... fully accelerated on all OS

[loscrossos]

I spent a good while repairing Zonos and enabling all possible accelerator libraries for CUDA Blackwell cards..

For this I fixed Bugs on Pytorch, brought improvements on Mamba, Causal Convid and what not...

Hybrid and Transformer models work at full speed on Linux and Windows. then i said.. what the heck.. lets throw MacOS into the mix... MacOS supports only Transformers.

did i mentioned, that the installation is ultra easy? like 5 copy paste commmands.

behold... core Zonos!

It will install Zonos on your PC fully working with all possible accelerators.

https://github.com/loscrossos/core_zonos

Step by step tutorial for the noob:

mac: https://youtu.be/4CdKKLSplYA

linux: https://youtu.be/jK8bdywa968

win: https://youtu.be/Aj18HEw4C9U

Check my other project to automatically setup your PC for AI development. Free and open source!:

https://github.com/loscrossos/crossos_setup

check out my reddit post

https://www.reddit.com/r/StableDiffusion/comments/1l0vtfk/so_i_repaired_zonos_woks_on_windows_linux_and/

[AznamirWoW]

Care to make 3.11 requirements?

[AznamirWoW]

Unfortunate the 3.12 fails

(venv) G:\core_zonos>test.py
Traceback (most recent call last):
File "G:\core_zonos\test.py", line 3, in
from zonos.model import Zonos
File "G:\core_zonos\zonos\model.py", line 11, in
from zonos.backbone import BACKBONES
File "G:\core_zonos\zonos\backbone_init_.py", line 4, in
from .mamba_ssm import MambaSSMZonosBackbone
File "G:\core_zonos\zonos\backbone_mamba_ssm.py", line 3, in
from mamba_ssm.models.mixer_seq_simple import create_block
File "G:\core_zonos\venv\Lib\site-packages\mamba_ssm_init.py", line 3, in
from mamba_ssm.ops.selective_scan_interface import selective_scan_fn, mamba_inner_fn
File "G:\core_zonos\venv\Lib\site-packages\mamba_ssm\ops\selective_scan_interface.py", line 16, in
from mamba_ssm.ops.triton.layer_norm import _layer_norm_fwd
File "G:\core_zonos\venv\Lib\site-packages\mamba_ssm\ops\triton\layer_norm.py", line 169, in
@triton.autotune(
^^^^^^^^^^^^^^^^
File "G:\core_zonos\venv\Lib\site-packages\triton\runtime\autotuner.py", line 378, in decorator
return Autotuner(fn, fn.arg_names, configs, key, reset_to_zero, restore_value, pre_hook=pre_hook,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "G:\core_zonos\venv\Lib\site-packages\triton\runtime\autotuner.py", line 130, in init
self.do_bench = driver.active.get_benchmarker()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "G:\core_zonos\venv\Lib\site-packages\triton\runtime\driver.py", line 23, in getattr
self._initialize_obj()
File "G:\core_zonos\venv\Lib\site-packages\triton\runtime\driver.py", line 20, in _initialize_obj
self._obj = self._init_fn()
^^^^^^^^^^^^^^^
File "G:\core_zonos\venv\Lib\site-packages\triton\runtime\driver.py", line 8, in _create_driver
raise RuntimeError(f"{len(actives)} active drivers ({actives}). There should only be one.")
RuntimeError: 0 active drivers ([]). There should only be one.

[AznamirWoW]

Sorry, seems that uv did not install proper torch and the error above is because of 2.7.0+cpu.

[jinno-kuma]

amazing thanks