-
Notifications
You must be signed in to change notification settings - Fork 116
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Convert the alignment model #5
Comments
This model was not tested yet, only Retinaface, ArcFace and GenderAge models from insightface model zoo are currently supported. This seems to be related to replacement of PRelu nodes with Relu, you can try commenting all PRelu/Relu related code in insight2onnx.py if you are using onnx 1.7.0 and trt > 7.0, since newer versions seems to support PRelu. |
@wuqiangch, I have fixed conversion in latest commit. Conversion from MXNet to TRT was successfull, but I haven't checked outputs, please test it and leave comment if it helped. BTW, also this update seems to normally convert newer SubCenter-ArcFace pretrained models. |
Closing, since there is no activity, and problem should have been solved in older commits. Feel free to reopen it if needed. |
@SthPhoenix Hi ,I used the code to convert the face alignment mode. https://github.com/deepinsight/insightface/tree/master/alignment/coordinateReg
but the code of checker.check_graph(graph) get the bellow error info:
onnx.onnx_cpp2py_export.checker.ValidationError: Nodes in a graph must be topologically sorted, however input 'reshape4' of node: input: "conv_1_relu_gamma" input: "reshape4" output: "slope4" name: "slope4" op_type: "Reshape" is not output of any previous nodes.
The text was updated successfully, but these errors were encountered: