Yolov5 tflite uint8 export

WiFiHan commented 2 years ago

Search before asking

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

YOLOv5 Component

No response

Bug

Yolov5 tflite model export with uint8 input doesn't work. I work in google colab. I want to export tflite-uint8 model.

The model is exported to tflite fp-16 when not adding --uint8, but when adding --uint8, no tflite model found in the export path. Instead, there's only 'saved_model.pb' file.

I checked #5707, and downgraded my flatbuffer library to 1.12.

Environment

-Google colab -Python 3.7.13

Minimal Reproducible Example

My code is:

%cd /content/drive/MyDrive/yolov5/ !python export.py --img 640 --conf 0.25 --weights /content/drive/MyDrive/new_yolov5/runs/train/20220523_1945003/weights/best.pt/ --int8 --include saved_model tflite

Additional

No response

Are you willing to submit a PR?

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

github-actions[bot] commented 2 years ago

👋 Hello @WiFiHan, 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.

WiFiHan commented 2 years ago

I've found that the code below works:

!python export.py --weights /content/drive/MyDrive/yolov5/runs/train/fine_tuning2/weights/best.pt --include saved_model tflite --int8

When data path is not specified, the code above worked. But still the code below doesn't export tflite model:

!python export.py --img 640 --data /content/drive/MyDrive/test5_mk_od/data/dataset.yaml --weights /content/drive/MyDrive/yolov5/runs/train/fine_tuning2/weights/best.pt --include saved_model tflite --int8

glenn-jocher commented 2 years ago

@WiFiHan --data is just used for quantization. Default --data is COCO128. You can pass your own if you want the same way as coco128.yaml.

WiFiHan commented 2 years ago

@WiFiHan --data is just used for quantization. Default --data is COCO128. You can pass your own if you want the same way as coco128.yaml.

But the problem is that when I pass my own data then quantization and tflite export aren't done. Do you mean there's no problem when I pass '--data 'default with my own custom weights(.pt file)?

glenn-jocher commented 2 years ago

@WiFiHan a dataset is always passed: https://github.com/ultralytics/yolov5/blob/27911dc824b28f6d805dc88360a3313ce0c9cae5/export.py#L569

If you have a reproducible bug please submit a bug report with exact code to reproduce (including a commonly available dataset, not your custom dataset).

WiFiHan commented 2 years ago

Thanks for your kind help! With more research, I'll submit a bug report with exact code.

WiFiHan commented 2 years ago

I found that if I pass my own custom dataset --data /content/drive/MyDrive/test5_mk_od/data/dataset.yaml , then the train data of dataset.yaml file must not be .txt; it worked when it was written in the form of directory.

names: [Book, Calculator, Mobile phone, Pencil case, Postit]
nc: 5
train: /content/drive/MyDrive/test5_mk_od/data/images/train #worked when directory was passed
val: /content/drive/MyDrive/test5_mk_od/data/images/val

names: [Book, Calculator, Mobile phone, Pencil case, Postit]
nc: 5
train: /content/drive/MyDrive/test5_mk_od/data/train.txt #not worked when txt file was passed
val: /content/drive/MyDrive/test5_mk_od/data/val.txt

github-actions[bot] commented 2 years ago

👋 Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 🚀 resources:

Wiki – https://github.com/ultralytics/yolov5/wiki
Tutorials – https://docs.ultralytics.com/yolov5
Docs – https://docs.ultralytics.com

Access additional Ultralytics ⚡ resources:

Ultralytics HUB – https://ultralytics.com/hub
Vision API – https://ultralytics.com/yolov5
About Us – https://ultralytics.com/about
Join Our Team – https://ultralytics.com/work
Contact Us – https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 🚀 and Vision AI ⭐!

glenn-jocher commented 1 year ago

@WiFiHan, great job on identifying the issue! Feel free to include this observation in your bug report. It will surely help the developers in resolving the issue promptly. If you have any further questions or need assistance with anything else, don't hesitate to ask!

ultralytics / yolov5