I am trying to install mlc-ai on aws amazon-linux machine

[vk-maurya]

No matching distribution found

I am trying to install mlc-ai dependency using the below command on my aws amazon-Linux machine. but it is saying No matching distribution found for mlc-ai-nightly-cu118.

command :

$ pip install --pre --force-reinstall mlc-ai-nightly-cu118 mlc-chat-nightly-cu118 -f https://mlc.ai/wheels
Error
Looking in links: https://mlc.ai/wheels
ERROR: Could not find a version that satisfies the requirement mlc-ai-nightly-cu118 (from versions: none)
ERROR: No matching distribution found for mlc-ai-nightly-cu118

Linux machine details :

$ cat /etc/*-release

NAME="Amazon Linux"
VERSION="2"
ID="amzn"
ID_LIKE="centos rhel fedora"
VERSION_ID="2"
PRETTY_NAME="Amazon Linux 2"
ANSI_COLOR="0;33"
CPE_NAME="cpe:2.3:o:amazon:amazon_linux:2"
HOME_URL="https://amazonlinux.com/"
Amazon Linux release 2 (Karoo)

Python version :

Python 3.10.12

Nvidia driver :

$ nvcc --version

nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Wed_Sep_21_10:33:58_PDT_2022
Cuda compilation tools, release 11.8, V11.8.89
Build cuda_11.8.r11.8/compiler.31833905_0

[junrushao]

This is weird. The prebuilt is based on standard manylinux convention so it’s supposed to be available. Could you download the specific wheel and try to install locally?

[vk-maurya]

Locally and Google Colab both have Ubuntu it is working there. However AWS Machine has Amazon Linux 2 which causing the issue.

[Hzfengsy]

Try to upgrade pip?

[vk-maurya]

I have upgraded the Latest pip version: pip 23.2.1. still not working

[junrushao]

How about:

python -m pip install ...

[linjie7674]

I have the same problem, and I tried to install it by downloading the wheel, but it still does not work.

Command

Linux machine details :

Python version:

Python 3.10.12

[Hzfengsy]

Is it an x86 machine or an arm one?

[junrushao]

Would love to have more info to debug. Could you guys use the command below:

python3 -m pip install -v …

[vk-maurya]

I am getting same error even I have tried other amazon linux machine :

[yzh119]

Hi @vk-maurya , our pip wheels are built upon manylinux_2_28 to install vulkan sdk (the glibc in manylinux_2014 is too old to install a relatively new version of vulkan sdk).

manylinux list distribution compatibility in their README, They didn't explicitly mention Amazon Linux 2 but I suppose it's only compatible with manylinux_2014, not manylinux_2_28.

We can compile standalone CUDA wheels based on manylinux_2014 without incorporating Vulkan SDK to make our wheels compatible with more distributions. @junrushao @tqchen WDYT?

[tqchen]

@yzh119 yah i think this is something we can do if that is not as complicated

[junrushao]

Some additional discussion around Amazon Linux can be found here: pypa/manylinux#1332

[qizzzh]

Hitting the same problem on GCP linux VMs.

[smpurkis]

I'm having similar issues with linux arm oracle cloud server.

[williamcastrillon]

Hey! I was having a similar issue, but I found this mlc-ai repo and I was able to solve the problem I was having!
Hope that helps!

[1748995976]

1. use pip debug --verbose to see related information
2. use personal computer download file in https://mlc.ai/wheels ,then upload it to your server computer
3. use python -m pip install ***.whl to install it

About this post, you can:

1. download mlc_ai_nightly_cu118-0.12.dev1689-cp310-cp310-manylinux_2_28_x86_64.whl
2. execute python -m pip install mlc_ai_nightly_cu118-0.12.dev1689-cp310-cp310-manylinux_2_28_x86_64.whl

[ameryzhu]

I have a problem of install too. I fixed it by upgrading conda python from 3.7 to 3.11.

Problem:
python3 -m pip install --pre -U -f https://mlc.ai/wheels mlc-ai-nightly --verbose
Using pip 24.0 from /Users/zhuzhe/miniconda3/envs/mlc/lib/python3.7/site-packages/pip (python 3.7)
Looking in links: https://mlc.ai/wheels
ERROR: Could not find a version that satisfies the requirement mlc-ai-nightly (from versions: none)
ERROR: No matching distribution found for mlc-ai-nightly

Problem Fixed:
conda install python=3.11