pipenv: Unable to install mmcv-full with pipenv.
Issue description
Unable to install mmcv-full with pipenv.
Expected result
Can be installed without problems
Actual result
An error will occur as shown below.
⋊> /m/n/k/c/work on dev-mmdet-2.23.0 ⨯ pipenv lock 12:10:07
Locking [dev-packages] dependencies...
Locking [packages] dependencies...
Building requirements...
Resolving dependencies...
✔ Success!
Updated Pipfile.lock (a7d644)!
⋊> /m/n/k/c/work on dev-mmdet-2.23.0 ⨯ pipenv sync 12:10:51
Installing dependencies from Pipfile.lock (a7d644)...
An error occurred while installing https://download.openmmlab.com/mmcv/dist/cu113/torch1.10.0/mmcv_full-1.4.7-cp38-cp38-manylinux1_x86_64.whl#egg=mmcv-full --hash=sha256:7459572935a7ba6293fee619c205583d76cd832464ffe9f15115ebe0396450e4! Will try again.
🐍 ▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉ 14/14 — 00:00:20
Installing initially failed dependencies...
[pipenv.exceptions.InstallError]: Collecting mmcv-full==1.4.7
[pipenv.exceptions.InstallError]: Using cached https://download.openmmlab.com/mmcv/dist/cu113/torch1.10.0/mmcv_full-1.4.7-cp38-cp38-manylinux1_x86_64.whl (47.7 MB)
[pipenv.exceptions.InstallError]: ERROR: THESE PACKAGES DO NOT MATCH THE HASHES FROM THE REQUIREMENTS FILE. If you have updated the package versions, please update the hashes. Otherwise, examine the package contents carefully; someone may have tampered with them.
[pipenv.exceptions.InstallError]: mmcv-full==1.4.7 from https://download.openmmlab.com/mmcv/dist/cu113/torch1.10.0/mmcv_full-1.4.7-cp38-cp38-manylinux1_x86_64.whl#egg=mmcv-full (from -r /tmp/pipenv-vznfd55f-requirements/pipenv-1z06yhcm-requirement.txt (line 1)):
[pipenv.exceptions.InstallError]: Expected sha256 7459572935a7ba6293fee619c205583d76cd832464ffe9f15115ebe0396450e4
[pipenv.exceptions.InstallError]: Got 95dde9139fec10789406aa13e4824c3187805f52f23acd5aa6b924094be5e427
ERROR: Couldn't install package: mmcv-full
Package installation failed...
☤ ▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉▉ 0/1 — 00:00:02
⋊> /m/n/k/c/work on dev-mmdet-2.23.0 ⨯
Steps to replicate
For commands only, it can be reproduced as follows.
pip install pipenv
pipenv --python 3.8.6
pipenv install "mmcv-full==1.4.7" -i https://download.openmmlab.com/mmcv/dist/cu113/t
orch1.10.0/index.html
Another method is to write it in the Pipfile in advance and then lock
and sync
.
[[source]]
url = "https://pypi.org/simple"
verify_ssl = true
name = "pypi"
[packages]
mmcv-full = {file = "https://download.openmmlab.com/mmcv/dist/cu113/torch1.10.0/mmcv_full-1.4.7-cp38-cp38-manylinux1_x86_64.whl"}
torch = {file = "https://download.pytorch.org/whl/cu113/torch-1.10.0%2Bcu113-cp38-cp38-linux_x86_64.whl"}
torchvision = {file = "https://download.pytorch.org/whl/cu113/torchvision-0.11.0%2Bcu113-cp38-cp38-linux_x86_64.whl"}
mmdet = "==2.23.0"
[dev-packages]
[requires]
python_version = "3.8"
After creating the above Pipfile, do pipenv lock
and pipenv sync
.
pipenv lock
pipenv sync
Environment
It is running in the following environment.
⋊> ~ nvidia-container-cli info 12:05:06
NVRM version: 465.19.01
CUDA version: 11.3
Device Index: 0
Device Minor: 0
Model: NVIDIA Tesla T4
Brand: Nvidia
GPU UUID: GPU-b2fd7cfe-e194-bdb7-3a2a-78a5d8528d38
Bus Location: 00000000:00:1e.0
Architecture: 7.5
⋊> ~ lsb_release -a 12:05:32
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 18.04.6 LTS
Release: 18.04
Codename: bionic
⋊> ~
About this issue
- Original URL
- State: closed
- Created 2 years ago
- Comments: 16 (9 by maintainers)
Thanks @Keiku – there should be a new release in an hour or so that will included the latest changes and the fix for the original issue. The
torchvision
version error difference remains a mystery, but I will keep it in mind if we get any other issue reports with the same error like that. May end up closing at some point soon based on your positive feedback. Thanks!@matteius In pipenv version
2022.4.9.dev0
, the below error occur withlock
. I’m not sure why this is happening.I’ll try the main branch for now.