how convert the ts output to onnx?

[MECHAAMER]

can you guide me how convert the tx output to onnx?

I try this:

import torch
import onnx
import torchvision

model = torch.jit.load("/....../weights/best.ts")

# Set  input shape of the model
 input_shape = (1, 3, 224, 224)
 ​
 # Export  PyTorch model to ONNX format
 torch.onnx.export(model, torch.randn(input_shape), 'output.onnx')

this is the output:

Inputs:
     #0: r0 defined in (%r0 : Tensor = prim::TupleConstruct(%388, %386, %387), scope: nets.nn.QAT:: # /home/at-am/.local/lib/python3.8/site-packages/torch/ao/nn/quantized/modules/functional_modules.py:212:12
 )  (type 'Tensor')
     #1: 405 defined in (%405 : Tensor = onnx::Concat[axis=0](%402, %404), scope: nets.nn.QAT:: # /home/at-am/datasets/VisDrone/Yolo-qat/nets/nn.py:50:17
 )  (type 'Tensor')
 Outputs:
     #0: 406 defined in (%406 : Tensor[] = onnx::SplitToSequence[axis=1](%r0, %405), scope: nets.nn.QAT:: # /home/at-am/datasets/VisDrone/Yolo-qat/nets/nn.py:50:17
 )  (type 'List[Tensor]')