Pytorch and tensorrt seems can't work together on some machine. We haven't find the solution, but you can try the two-step convertion way: uninstall tensorrt if you have installed (pip uninstall tensorrt) convert the model to onnx use torch2onnx install tensorrt (pip install tensort_xxx...
Run demo follow https://github.com/open-mmlab/mmdeploy/blob/master/docs/en/02-how-to-run/convert_model.md python tools/deploy.py \ configs/mmdet/detection/detection_tensorrt_dynamic-320x320-1344x1344.py \ /root/mmdetection/configs/yolo/y...
I try to convert a simple network with QuantizeLinear/DequantizeLinear but it fails. The network is very simple that it only contains one tensor as input. Its definition could be found in test of pytorch (https://github.com/pytorch/pytorch/blob/master/test/onnx/model_defs/op_test.py). c...
🐛 Describe the bug nn.RNNBase.flatten_parameters function should be a no-op for export. Otherwise, export()/dynamo_export fail inside it with this error: File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/rnn.py", line 196, i...
Hello I'm trying to convert CLIP_Surgery model to onnx format. I stuck with the next error: Error: Fail: [ONNXRuntimeError] : 1 : FAIL : Load model from cs_vit_b_16_vis.onnx failed:vector::_M_range_check: __n (which is 1) >= this->size()...
Tensors and Dynamic neural networks in Python with strong GPU acceleration - symbolic_convert: Don't fail when we hit a undefined name · pytorch/pytorch@8c2aa0c
result = doc_converter.convert(input_doc_path) result.document.save_as_html(Path("example.html"),formula_to_mathml=True) result.document.save_as_markdown(Path("example.md")) if __name__ == "__main__": main() However, I got a "Formula not decoded" in the html and a "" in ...
I firstly convert pytorch nodel to onnx model Then I do onnx optimization with onnxoptimizer.optimize Then I do the inference on the optimized onnx model. But I also did the inference with the onnx model before optimization. Both gave me the aforementioned error. I guess it's because th...
(gm, example_inputs) File "D:\xu_git\dnnl_cb\pytorch\test\inductor\test_torchinductor.py", line 422, in compile_fx_wrapper return compile_fx(model_, example_inputs_) File "C:\Users\Xuhan\.conda\envs\win_mkl_static\lib\site-packages\torch\_inductor\compile_fx.py", line 1521, in...
Hi team! We also observed some issues when compiling vision models(swin and donut) with batch_size > 1 after bumping to Neuron SDK 2.21.0: *** Compiling tiny-random-SwinModel *** Using Neuron: --optlevel 2 ..root = neuronxcc/starfish...