pinto0309 / facemesh_onnx_tensorrt Goto Github PK

hello,

do you think it would be possible ot run these onnx models on nvidia deepstream using nvinfer ?
i think it can read onnx, but i don't know enough if this would work.

thanks,
tom

Hi Pinto, i would like to test your programme, but i think i have an error with pre-trained models because i am not rebuilding ther on my plateform.
I work with a JetsonNX with jetpack 4.4 (i can use 4.6 i that needed), with cuda 10.2, cudnn 8.0, onnx==1.11.0, onnxgraphsurgeon==0.3.9, onnx-tensorrt==7.2.2.3.0, onnxruntime==1.8.0

My error is at yolov4_headdetection_480x640_post.onnx:

onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Load model from yolov4_headdetection_480x640_post.onnx failed:/mnt/vss/_work/1/s/onnxruntime/core/graph/model.cc:111 onnxruntime::Model::Model(onnx::ModelProto&&, const PathString&, const IOnnxRuntimeOpSchemaRegistryList*, const onnxruntime::logging::Logger&) Unknown model file format version.

When i change my model by yolov4 headdetection 480x640.onnx, i have an error on the next model face_mesh_Nx3x192x192 _post.onnx :

onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Load model from face_mesh_Nx3x192x192_post.onnx failed:/mnt/vss/_work/1/s/onnxruntime/core/graph/model.cc:111 onnxruntime::Model::Model(onnx::ModelProto&&, const PathString&, const IOnnxRuntimeOpSchemaRegistryList*, const onnxruntime::logging::Logger&) Unknown model file format version.

If i'm understand well, i must convert my model to post format model, but i don't really understand how i can do that.
I read the "facemesh_postprocess.py", but i'm don't understand how i can use it.

Is it possible for you, to explain how i can obtain these post models converted on my platform?
Or what version of onnxruntime i must have ?

Thank you for your answer :)

Do you have any plan to transfer mediapipe blendshape model to onnx/trt?

ref link: https://storage.googleapis.com/mediapipe-assets/Model%20Card%20Blendshape%20V2.pdf

error
read_config_from_model)
onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Load model from yolov4_headdetection_480x640_post.onnx failed:D:\3\s\onnxruntime\core\graph\model.cc:110 onnxruntime::Model::Model Unknown model file format version.