-
-
Notifications
You must be signed in to change notification settings - Fork 16.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Problems when downloading requirements #2047
Comments
👋 Hello @SebaSilvaS, thank you for your interest in 🚀 YOLOv5! Please visit our ⭐️ Tutorials to get started, where you can find quickstart guides for simple tasks like Custom Data Training all the way to advanced concepts like Hyperparameter Evolution. If this is a 🐛 Bug Report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you. If this is a custom training ❓ Question, please provide as much information as possible, including dataset images, training logs, screenshots, and a public link to online W&B logging if available. For business inquiries or professional support requests please visit https://www.ultralytics.com or email Glenn Jocher at glenn.jocher@ultralytics.com. RequirementsPython 3.8 or later with all requirements.txt dependencies installed, including $ pip install -r requirements.txt EnvironmentsYOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):
StatusIf this badge is green, all YOLOv5 GitHub Actions Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training (train.py), testing (test.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu every 24 hours and on every commit. |
To add to this: I also stumbled across this issue late last night, but not using conda. I cloned the project, and tried running Attempting to run the train script failed as above on the SciPy import, so I ran This allowed me to slowly import all modules, INCLUDING TORCHVISION, which by using straight After all imports are done, attempting to train fails on checking requirements as the requirement checking logic appears to deem torchvision==0.8.2 as invalid. |
@SebaSilvaS @setho246 it appears you may have environment problems. Please ensure you meet all dependency requirements if you are attempting to run YOLOv5 locally. If in doubt, create a new virtual Python 3.8 environment, clone the latest repo (code changes daily), and RequirementsPython 3.8 or later with all requirements.txt dependencies installed, including $ pip install -r requirements.txt EnvironmentsYOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):
StatusIf this badge is green, all YOLOv5 GitHub Actions Continuous Integration (CI) tests are passing. These tests evaluate proper operation of basic YOLOv5 functionality, including training (train.py), testing (test.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu. |
@SebaSilvaS @setho246 also note your particular environment may require custom installation commands for torch and torchvision. See https://pytorch.org/get-started/locally/ for specifics. You would run these custom torch installation commands prior to pip install -r requirements.txt. |
Cheers, I'll test some more when I get the chance and report back |
I realized I had CUDA 10.1, so the latest version I could download for torchvision was 0.2.2.post3. However, this update does not fix my problem. Here is my environment, it is running on python 3.8.5:
EDITED: I just downgraded torchvision to 0.8.1 and, therefore, pytorch to 1.7.0. The problem persists. |
I got a "patch" solution. I change the "check_requirements" function to:
Since pycocotools is not needed in train.py nor detect.py, now I'm being able to train with a custom dataset. The only thing that I'm still messing around with is the wandb 1,2,3 option... when I press 2, it asks for the 40 digits code which I'm not being able to paste (not even write by hand). |
@SebaSilvaS thanks for the feedback. What environment are you having difficulty with wandb login? @AyushExel might be able to help you with that, he's our resident wandb expert :) |
@SebaSilvaS what environment are you using? I tried authenticating wandb in colab, jupyter and a local terminal and I was able to login without problems. Let me know how I can reproduce your problem. |
@glenn-jocher and @AyushExel, thanks for your time and attention. Im running it on a local environment, running on Python 3.8.5:
I'm following this steps:
Even if I want to write text manually when my keycode is asked I can't. I haven't tried updating the repo since yesterday though... should I try?
|
@SebaSilvaS ctrl+V wont work on cmd. And when writing manually it wont show what you've written as it's treated as a secure password. Just copy the code and then right click on the cmd prompt and then click on paste. That should work |
It worked! I just had to righ click. Even though nothing happened when I pressed enter, the code was right and no error raised. |
have you solved it? I met same problem |
🐛 Bug
First of all, thanks a lot to the people behind this amazing repo. I'm quite new in GitHub, so please understand if I make common newbie mistakes when asking or raising this issue.
Having mentioned the above, I'm having problems when installing this repo's dependencies. Particularly with "torch==0.8.1", "scikit-learn==0.19.2", and "pycocotools>=2.0". The error throws a huge red message (I'm not sure if I can copy it completely here).
To Reproduce (REQUIRED)
I'm using none extra code at all. The steps I followed are the following:
8.a. open the requirements.txt file and erase the "#" from the
8.b. copy and paste the "pip install -r requirements.txt" command from this repo
ERROR: No matching distribution found for torchvision>=0.8.1" is displayed in the screen.
So, I decided to install torchvision using the following command: "conda install pytorch torchvision torchaudio cudatoolkit=10.1 -c pytorch", gathered from https://pytorch.org/get-started/locally/.
Next, I ran the "pip install -r requirements.txt" command again and it started doing it thing. Here is where I get the huge red error text.
Output:
Expected behavior
I would expect every dependencies to be correctly and straight-forwardly installed.
Environment
If applicable, add screenshots to help explain your problem.
Additional context
The text was updated successfully, but these errors were encountered: