Cannot Load best.pt Without Full YOLOv5 Source Code

[dntcoder]

Search before asking

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

Question

ModuleNotFoundError: No module named 'models'
I encountered an issue when deploying my custom-trained YOLOv5 model (best.pt). Loading the model file depends on the full YOLOv5 source code structure, which makes it inconvenient for deployment in production or lightweight services. Is there a better way to load the model directly in an inference script without relying on the entire YOLOv5 codebase?

Additional

No response

[UltralyticsAssistant]

👋 Hello @dntcoder, 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 a minimum reproducible example to help us debug it.

If this is a custom training ❓ Question, please provide as much information as possible, including dataset image examples and training logs, and verify you are following our Tips for Best Training Results.

Requirements

Python>=3.8.0 with all requirements.txt installed including PyTorch>=1.8. 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):

• 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, validation, inference, export and benchmarks on macOS, Windows, and Ubuntu every 24 hours and on every commit.

This is an automated response. An Ultralytics engineer will assist you here soon 😊.

[pderrenger]

For deployment without the full YOLOv5 source code, export your model to ONNX format using python export.py --weights best.pt --include onnx then load it with OpenCV DNN or ONNX Runtime, or alternatively install the ultralytics package and load your model with from ultralytics import YOLO; model = YOLO('best.pt') which handles dependencies automatically.

[MohiteYash]

I also encountered this issue when trying to deploy a custom-trained best.pt model in a minimal environment. While exporting to ONNX or using the ultralytics package works well, I found it helpful to create a quick and lightweight script using torch.hub.load() for basic inference without needing the entire repo or installation.

It’s especially useful for fast prototyping or sharing with beginners who just want to run predictions on their images with minimal setup.

[pderrenger]

Great point @MohiteYash! The torch.hub.load('ultralytics/yolov5', 'custom', 'best.pt') approach is indeed perfect for quick deployment scenarios where you want to avoid package installations while still leveraging PyTorch Hub's automatic dependency handling.