it seems that check_wandb_resume don't support multiple input files of images.

[murphypei]

it's just a minor bug about compatibility

🐛 Bug

for form2) or 3) in dataset yaml

# train and val data as 1) directory: path/images/, 2) file: path/images.txt, or 3) list: [path1/images/, path2/images/]

The data_dict['train'] and data_dict['val'] are list in here.

if data_dict['train'].startswith(WANDB_ARTIFACT_PREFIX):
    ...

if data_dict['val'].startswith(WANDB_ARTIFACT_PREFIX):
   ...

To Reproduce (REQUIRED)

Input:

train:
  [
    /path/image_list1.txt,
    /path/image_list2.txt,
    /path/image_list3.txt
  ]

Output:

YOLOv5 🚀 v4.0-185-g74276d5 torch 1.7.1+cu101 CUDA:0 (GeForce RTX 2080 Ti, 10989.4375MB)
                                             CUDA:1 (GeForce RTX 2080 Ti, 10989.4375MB)
                                             CUDA:2 (GeForce RTX 2080 Ti, 10989.4375MB)
                                             CUDA:3 (GeForce RTX 2080 Ti, 10989.4375MB)

    if data_dict['train'].startswith(WANDB_ARTIFACT_PREFIX):
AttributeError: 'list' object has no attribute 'startswith'

Environment

If applicable, add screenshots to help explain your problem.

• OS: [Ubuntu]
• GPU [2080 Ti]

[github-actions]

👋 Hello @murphypei, 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.

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 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), testing (test.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu every 24 hours and on every commit.

[glenn-jocher]

@murphypei thanks for the bug report!

@AyushExel this is true, the dataset yaml has 3 options for specifying train and val sets. The base values can be either 1) a *.txt file of image paths, or 2) a directory of images. Then 1 or 2 can be used by themselves (just one path, just one directory), or in a list (multiple *.txt files, multiple directories, or both).

This makes it easy for organizations to train on multiple datasets or multiple collects without having to merge the underlying data, i.e.:

train: [path/to/dir1, path/to/dir2, path/to/dir3, ...]

[AyushExel]

@glenn-jocher @murphypei thanks for reporting this. I didn't know about the 3rd format of the dataset, might've missed it. I'll fix this right now.

[AyushExel]

@glenn-jocher #2719 Should fix this issue. I tested it on the following dataset format and all operations including resume worked fine.

train: [../coco128/images/train2017/ , ../coco128/images/train20172/] # 128 images
val: [../coco128/images/train2017/ , ../coco128/images/train20172/] # 128 images

@murphypei let me know if the problem still exists after this is merged

[glenn-jocher]

@AyushExel thanks for the quick fix! I've merged PR #2719 now.

@murphypei please git pull to receive this update and let us know if you have any other issues.

[murphypei]

@AyushExel @glenn-jocher thanks, I have tested it without any issues.