Can't reproduce the outputs from the detect.py file on images/bus or images/zidane locally

[RobinBram]

Hi!

Having real trouble with this repository. Have tried training on custom data, everyhing seems fine but the model does not do any progress. I can't either get the detect.py to detect any objects in the image when following the code from the Tutorial at https://colab.research.google.com/github/ultralytics/yolov5/blob/master/tutorial.ipynb#scrollTo=zR9ZbuQCH7FX . Here are some outputs from my notebook:
Setup complete. Using torch 1.7.0 _CudaDeviceProperties(name='GeForce GTX 1650', major=7, minor=5, total_memory=4096MB, multi_processor_count=14)

And when running !python detect.py --weights yolov5l.pt --img 640 --conf 0.25 --source data/images/

Namespace(agnostic_nms=False, augment=False, classes=None, conf_thres=0.25, device='', exist_ok=False, img_size=640, iou_thres=0.45, name='exp', project='runs/detect', save_conf=False, save_txt=False, source='data/images/', update=False, view_img=False, weights=['yolov5l.pt'])
Downloading https://github.com/ultralytics/yolov5/releases/download/v4.0/yolov5l.pt to yolov5l.pt...

Fusing layers...
image 1/2 C:\Users\robr\Documents\Python Scripts\yolov5\data\images\bus.jpg: 640x480 Done. (0.227s)
image 2/2 C:\Users\robr\Documents\Python Scripts\yolov5\data\images\zidane.jpg: 384x640 Done. (0.158s)
Results saved to runs\detect\exp6
Done. (0.454s)
YOLOv5 v4.0-53-g2acbe96 torch 1.7.0 CUDA:0 (GeForce GTX 1650, 4096.0MB)

0%| | 0.00/90.2M [00:00<?, ?B/s]
2%|1 | 1.44M/90.2M [00:00<00:06, 15.0MB/s]
5%|5 | 4.62M/90.2M [00:00<00:03, 25.8MB/s]
9%|9 | 8.52M/90.2M [00:00<00:02, 32.6MB/s]
16%|#5 | 14.4M/90.2M [00:00<00:01, 43.1MB/s]
27%|##6 | 24.3M/90.2M [00:00<00:01, 62.8MB/s]
37%|###6 | 33.1M/90.2M [00:00<00:00, 72.4MB/s]
44%|####4 | 40.0M/90.2M [00:00<00:00, 68.1MB/s]
52%|#####1 | 46.5M/90.2M [00:00<00:00, 62.3MB/s]
58%|#####8 | 52.5M/90.2M [00:00<00:00, 61.4MB/s]
65%|######4 | 58.4M/90.2M [00:01<00:00, 61.5MB/s]
71%|#######1 | 64.3M/90.2M [00:01<00:00, 56.4MB/s]
77%|#######7 | 69.8M/90.2M [00:01<00:00, 56.9MB/s]
84%|########3 | 75.5M/90.2M [00:01<00:00, 57.1MB/s]
90%|########9 | 81.1M/90.2M [00:01<00:00, 57.3MB/s]
96%|#########5| 86.6M/90.2M [00:01<00:00, 57.4MB/s]
100%|##########| 90.2M/90.2M [00:01<00:00, 56.4MB/s]
Model Summary: 392 layers, 47025981 parameters, 0 gradients, 115.4 GFLOPS.

I have tried all the different pretrained versions of yolo. Is this a version issue in some way and what can I do to get this to work?

[github-actions]

👋 Hello @RobinBram, 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 [email protected].

Requirements

Python 3.8 or later with all requirements.txt dependencies installed, including torch>=1.7. To install run:

$ pip install -r requirements.txt

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 Notebook with free GPU:
• Kaggle Notebook with free GPU: https://www.kaggle.com/ultralytics/yolov5
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Docker Image https://hub.docker.com/r/ultralytics/yolov5. 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), testing (test.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu every 24 hours and on every commit.

[RobinBram]

When creating a new environment and installing pytorch for cpu, it works fine!

[glenn-jocher]

May be a windows or environment issue. I'll paste our default environmental issue below:

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 pip install -r requirements.txt again. We also highly recommend using one of our verified environments below.

Requirements

Python 3.8 or later with all requirements.txt dependencies installed, including torch>=1.7. To install run:

$ pip install -r requirements.txt

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 Notebook with free GPU:
• Kaggle Notebook with free GPU: https://www.kaggle.com/ultralytics/yolov5
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Docker Image https://hub.docker.com/r/ultralytics/yolov5. See Docker Quickstart Guide

Status

If 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.

[cloudwsx]

maybe it is a CUDA problem.
#1910

[RobinBram]

It works now. Degraded CUDA to 10.1 and reinstalled everything

[RobinBram]

Here are some training runs I did with the "bug".

[devpawann]

Degraded CUDA to 10.1

How do you exactly do that?
It's not working for me