error during the package installation
surendraphd opened this issue · 7 comments
Hi,
I encountered an error during the installation of the package using the setup.sh script, specifically with the causal-conv1d and mamba-ssm packages. Manual installation of these packages also results in the following error:
(saturn) user@user:~/Downloads/saturn$ pip install causal-conv1d>=1.2.0
Collecting mamba-ssm
Using cached mamba_ssm-1.2.2.tar.gz (35 kB)
Preparing metadata (setup.py) ... error
error: subprocess-exited-with-error
× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [11 lines of output]
Traceback (most recent call last):
File "", line 2, in
File "", line 34, in
File "/tmp/pip-install-5cf7j5zt/mamba-ssm_32f156da05014f5ea350e15e9a64e1fe/setup.py", line 103, in
raise RuntimeError(
RuntimeError: mamba_ssm is only supported on CUDA 11.6 and above. Note: make sure nvcc has a supported version by running nvcc -V.
torch.__version__ = 1.12.1+cu113
[end of output]
note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed
× Encountered error while generating package metadata.
╰─> See above for output.
note: This is an issue with the package mentioned above, not pip.
hint: See above for details.
(saturn) user@user:~/Downloads/saturn$ pip install mamba-ssm
Collecting mamba-ssm
Using cached mamba_ssm-1.2.2.tar.gz (35 kB)
Preparing metadata (setup.py) ... error
error: subprocess-exited-with-error
× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [11 lines of output]
Traceback (most recent call last):
File "", line 2, in
File "", line 34, in
File "/tmp/pip-install-iwpc6s9g/mamba-ssm_164c385c52324c8ebe2ef8f9265db92e/setup.py", line 103, in
raise RuntimeError(
RuntimeError: mamba_ssm is only supported on CUDA 11.6 and above. Note: make sure nvcc has a supported version by running nvcc -V.
torch.__version__ = 1.12.1+cu113
[end of output]
note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed
× Encountered error while generating package metadata.
╰─> See above for output.
note: This is an issue with the package mentioned above, not pip.
hint: See above for details.
I have also tried verifying my PyTorch version and CUDA compatibility, which appear to be correct:
Python 3.10.13 | packaged by conda-forge | (main, Dec 23 2023, 15:36:39) [GCC 12.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
import torch
print(f"PyTorch version: {torch.version}")
PyTorch version: 1.12.1+cu113
exit()
Could you please provide guidance on how to successfully install these packages?
Thanks
Hi @surendraphd,
Sorry for the error. I have pushed a change to setup.py. Would you be able to try re-installing?
Hi GuoJeff,
Thank you for your quick response. Unfortunately, the changes did not resolve the issue.
I am sharing the terminal log in the attached *.txt file for further review.
Thanks,
Hi @surendraphd,
Could you share your system configuration? I just re-tested installing the environment on 2 systems (with NVIDIA RTX 3090 and RTX A6000).
From the error in the text file: ImportError: cannot import name 'packaging' from 'pkg_resources', could you try pip install setuptools before installing the causal-conv1d and mamba-ssm packages?
Hi GuoJeff,
Thank you for your response.
Here is my system details.
I am using Ubuntu 20.04 with NVIDIA GeForce RTX 3070.
(base) user@user:~$ cat /etc/os-release
NAME="Ubuntu"
VERSION="20.04.6 LTS (Focal Fossa)"
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME="Ubuntu 20.04.6 LTS"
VERSION_ID="20.04"
VERSION_CODENAME=focal
UBUNTU_CODENAME=focal
(base) user@user:~$ lscpu
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
Address sizes: 46 bits physical, 48 bits virtual
CPU(s): 36
On-line CPU(s) list: 0-35
Thread(s) per core: 2
Core(s) per socket: 18
Socket(s): 1
NUMA node(s): 1
Vendor ID: GenuineIntel
CPU family: 6
Model: 85
Model name: Intel(R) Xeon(R) W-2295 CPU @ 3.00GHz
Stepping: 7
CPU MHz: 3000.000
CPU max MHz: 4800.0000
CPU min MHz: 1200.0000
BogoMIPS: 6000.00
Virtualization: VT-x
L1d cache: 576 KiB
L1i cache: 576 KiB
L2 cache: 18 MiB
L3 cache: 24.8 MiB
NUMA node0 CPU(s): 0-35
Hi @surendraphd,
I checked the Ubuntu system (on the 2 machines I have access to) and they have the same version as yours. I was able to install the environment with the latest changes which enforced the causal-conv1d and mamba-ssm versions.
Would you be able to try pip install setuptools and then manually try the pip commands for causal-conv1d and mamba-ssm again?
Hi GuaJeff,
Thank you for your support.
Although none of the suggested solutions worked for me initially, I managed to install the package successfully. If others face similar issues, they might find the following approach helpful:
-
Install all required packages except PyTorch with CUDA 11.3, causal-conv1d, and mamba-ssm.
-
After completing the initial installations, proceed with the following commands:
pip install torch==2.1.0 torchvision==0.16.0 torchaudio==2.1.0 --index-url https://download.pytorch.org/whl/cu118
pip install causal_conv1d-1.1.3.post1+cu118torch2.1cxx11abiFALSE-cp310-cp310-linux_x86_64.whl
pip install mamba_ssm-1.1.1+cu118torch2.1cxx11abiFALSE-cp310-cp310-linux_x86_64.whl
This method worked for me, allowing me to successfully run config.json located in the part-1 folder.
PS: Please note that my machine has CUDA 11.3 installed, and I believe PyTorch with CUDA 11.8 is also compatible with CUDA 11.3.
Thanks,
Hi @surendraphd,
Thank you for sharing your solution. I have updated the README to point to this issue in case others experience a similar problem.
Thank you also for trying out Saturn.