why change from python3.6 to python3.7?

[fei4xu]

Search before asking

• I have searched the YOLOv5 issues and discussions and found no similar questions.

Question

Hi,

I have a Jetson Nano with has python 3.6 and its pytorch version for python3.6.

I checked the readme.md and found that the required python version was changed from 3.6 to 3.7 recently ( commit ff8646c 2022/1/26) and then released as v6.1 tag. I didn't see any python or library upgrade around those commits.

I'd like to check if this is a planned version upgrade or python3.6 is still supported.

Thanks,

Fei

Additional

No response

[github-actions]

👋 Hello @fei4xu, 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://ultralytics.com or email support@ultralytics.com.

Requirements

Python>=3.7.0 with all requirements.txt installed including PyTorch>=1.7. To get started:

git clone https://github.com/ultralytics/yolov5  # clone
cd yolov5
pip install -r requirements.txt  # install

Environments

YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

• Google Colab and Kaggle notebooks with free GPU:
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Amazon Deep Learning AMI. See AWS Quickstart Guide
• Docker Image. See Docker Quickstart Guide

Status

If 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), validation (val.py), inference (detect.py) and export (export.py) on macOS, Windows, and Ubuntu every 24 hours and on every commit.

[glenn-jocher]

@fei4xu Python 3.6 has reached end of life and is no longer officially supported: https://endoflife.date/python

[fei4xu]

Thanks, so I can safely guess that there isn't a technical reason (like library dependency) to drop python 3.6 support.

Actually I just tried, on Jetson Nano (python 3.6.9), with the latest code on master branch, if I ignore this change (a19406b) and change opencv from 4.1.2 to 4.1.1 which comes default in Jetson, everything works just fine!

Python version itself and the dependent libraries' version are alwasy a headache for new users. I don't want to let new users like me spend lots of time upgrading python versions. so maybe we can still leave a note on the readme page that python3.6 still works but not recommanded?

Thanks again for the great work!

[glenn-jocher]

@fei4xu we can try to revert cv2 dependency to 4.1.1, but we can no longer state support for 3.6 as we no longer test for it.

[glenn-jocher]

@fei4xu good news 😃! Your original issue may now be fixed ✅ in PR #7645. This PR drops cv2 requirements to >=4.1.1. I think the second change that would be needed for native Jetson support would be to do a soft test of python version (display a warning) rather than a hard test.

To receive this update:

• Git – git pull from within your yolov5/ directory or git clone https://github.com/ultralytics/yolov5 again
• PyTorch Hub – Force-reload model = torch.hub.load('ultralytics/yolov5', 'yolov5s', force_reload=True)
• Notebooks – View updated notebooks
• Docker – sudo docker pull ultralytics/yolov5:latest to update your image

Thank you for spotting this issue and informing us of the problem. Please let us know if this update resolves the issue for you, and feel free to inform us of any other issues you discover or feature requests that come to mind. Happy trainings with YOLOv5 🚀!

[glenn-jocher]

@fei4xu unofficial 3.6.9 support seems to work across most operating systems (macos and ubuntu). See https://github.com/ultralytics/yolov5/runs/6234184629?check_suite_focus=true

[StephenBeirlaen]

Hi all, sorry for reviving this, but it might help the OP.
I can confirm that Yolov5 is now running with this PR on my Jetson Nano 2GB. Thanks! :)
I used this image to have all system dependencies installed for me. I also had to run a pip3 install --upgrade pip before I could install Yolov5 requirements.

Question though: I had to add a swapfile to make inference run because Yolov5 consumes a lot of memory (1.9/2 GB RAM and 2.1 GB of swap). Are these numbers normal for Yolov5? I am only interested in running inference on this device, no training. Anything I can tune to fit in 2GB RAM? Already switched to lightdm which saves me ~500 MB.
Thanks in advance!

[glenn-jocher]

@StephenBeirlaen great, thanks for the confirmation! Most of our focus has been on efficient resource utilization during training, i.e. reducing CUDA memory requirements, and not especially on edge device inference, so I can't provide you specifics there.

But that said, torch is primarily a training framework and you can definitely realize performance gains by exporting to TRT for CUDA deployments, i.e.:

Format	export.py --include	Model
PyTorch	-	yolov5s.pt
TorchScript	torchscript	yolov5s.torchscript
ONNX	onnx	yolov5s.onnx
OpenVINO	openvino	yolov5s_openvino_model/
TensorRT	engine	yolov5s.engine
CoreML	coreml	yolov5s.mlmodel
TensorFlow SavedModel	saved_model	yolov5s_saved_model/
TensorFlow GraphDef	pb	yolov5s.pb
TensorFlow Lite	tflite	yolov5s.tflite
TensorFlow Edge TPU	edgetpu	yolov5s_edgetpu.tflite
TensorFlow.js	tfjs	yolov5s_web_model/

Usage:

git clone https://github.com/ultralytics/yolov5 -b update/bench_gpu  # clone
cd yolov5
pip install -qr requirements.txt coremltools onnx onnxruntime-gpu openvino-dev  # install
pip install -U nvidia-tensorrt --index-url https://pypi.ngc.nvidia.com  # TensorRT

python utils/benchmarks.py --weights yolov5s.pt --img 640 --device 0

Colab Pro+ V100 High-RAM Results

benchmarks: weights=/content/yolov5/yolov5s.pt, imgsz=640, batch_size=1, data=/content/yolov5/data/coco128.yaml, device=0, half=False, test=False
Checking setup...
YOLOv5 🚀 v6.1-135-g7926afc torch 1.10.0+cu111 CUDA:0 (Tesla V100-SXM2-16GB, 16160MiB)
Setup complete ✅ (8 CPUs, 51.0 GB RAM, 46.7/166.8 GB disk)

Benchmarks complete (458.07s)
                   Format  mAP@0.5:0.95  Inference time (ms)
0                 PyTorch        0.4623                10.19
1             TorchScript        0.4623                 6.85
2                    ONNX        0.4623                14.63
3                OpenVINO           NaN                  NaN
4                TensorRT        0.4617                 1.89
5                  CoreML           NaN                  NaN
6   TensorFlow SavedModel        0.4623                21.28
7     TensorFlow GraphDef        0.4623                21.22
8         TensorFlow Lite           NaN                  NaN
9     TensorFlow Edge TPU           NaN                  NaN
10          TensorFlow.js           NaN                  NaN