search

microsoft / MMdnn

MMdnn is a set of tools to help users inter-operate among different deep learning frameworks. E.g. model conversion and visualization. Convert models between Caffe, Keras, MXNet, Tensorflow, CNTK, PyTorch Onnx and CoreML.
MIT License
5.79k stars 964 forks source link

Convert trained yolov3 Darknet-53 custom model to tensorflow model #832

Open AICodePanda opened 4 years ago

AICodePanda commented 4 years ago

Platform: ubuntu 16.04 Python version: 3.6 Source framework: mxnet (gluon-cv) Destination framework: Tensorflow Model Type: Object detection Pre-trained model path: https://gluon-cv.mxnet.io/build/examples_detection/demo_yolo.html#sphx-glr-build-examples-detection-demo-yolo-py

I have trained a custom yolov3 Darknet-53 model (yolo3_darknet53_custom.params) using gluon-cv (mxnet). I need to convert the yolo3_darknet53_custom.params (mxnet) model to yolo3_darknet53_custom.pb (tensorflow)

Also, I see https://pypi.org/project/mmdnn/ object detection is under on-going Models.

Queries: Does mmdnn supports yolo models (object detection) conversion trained using gluon-cv (mxnet) in general? Is there any way or work around by which I can convert yolo models?

Any leads would be great! Thank you

XiaoXYe commented 4 years ago

Hi @rohit167 , we haven't test MMdnn with yolo models, so there might be some unsupported ops and accuracy loss.