Cannot use NMS with edgetpu export

democat3457 commented 2 years ago

Search before asking

[X] I have searched the YOLOv5 issues and found no similar bug report.

YOLOv5 Component

Export

Bug

Trying to export a PyTorch model as an edgetpu tflite model with the NMS flag enabled results in this error when exporting:

Edge TPU: starting export with Edge TPU compiler 16.0.384591198...
Edge TPU Compiler version 16.0.384591198
Started a compilation timeout timer of 180 seconds.
ERROR: Regular TensorFlow ops are not supported by this interpreter. Make sure you apply/link the Flex delegate before inference.
ERROR: Node number 270 (FlexCombinedNonMaxSuppression) failed to prepare.

Compilation failed: Model failed in Tflite interpreter. Please ensure model can be loaded/run in Tflite interpreter.
Compilation child process completed within timeout period.
Compilation failed! 

Edge TPU: export failure: Command '['edgetpu_compiler', '-s', '-o', '/path/to/weights', '/path/to/weights/best-int8.tflite']' returned non-zero exit status 1.

Everything before the Edge TPU export step, including the Tensorflow Lite export, succeeds just fine.

This seems similar to #6799

Environment

YOLOv5 🚀 v6.1-336-gd7411ad Python-3.8.10 torch-1.11.0+cu102 CPU
OS: Ubuntu 20.04
Python 3.8.10

Minimal Reproducible Example

python export.py --include edgetpu --nms

Additional

No response

Are you willing to submit a PR?

[X] Yes I'd like to help by submitting a PR!

github-actions[bot] commented 2 years ago

👋 Hello @democat3457, 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 commented 2 years ago

@democat3457 as the error says the NMS TF ops are not supported for Edge TPU delegates.

You can export Edge TPU models using the standard arguments and then follow the Usage instructions after export:

Detect:          python detect.py --weights yolov5s-int8.tflite 
Validate:        python val.py --weights yolov5s-int8.tflite 
PyTorch Hub:     model = torch.hub.load('ultralytics/yolov5', 'custom', 'yolov5s-int8.tflite')

democat3457 commented 2 years ago

Ah rip, ok.

mhaghighat commented 1 year ago

@democat3457 Did you find a workaround for this?

democat3457 commented 1 year ago

@democat3457 Did you find a workaround for this?

No, I just ended up implementing my own NMS using pytorch's.

ultralytics / yolov5