lzcchl

Results 19 comments of lzcchl

How can I adjust the position of quantization nodes to reduce data conversion?

![Image](https://github.com/user-attachments/assets/f8f6bd74-f2f4-405f-919a-7f583b754dd3)

How can I adjust the position of quantization nodes to reduce data conversion?

Through observing the automatic and manual placement of QDQ nodes in YOLOv7, I roughly understand that the parameters within the red box below need to be set equal. However, this...

How can I adjust the position of quantization nodes to reduce data conversion?

trt svg is: ![Image](https://github.com/user-attachments/assets/c2ab3e75-3b28-4565-8d9a-0eb8364533ee)

How can I adjust the position of quantization nodes to reduce data conversion?

onnx graph is: ![Image](https://github.com/user-attachments/assets/5c420ddd-f92e-4af3-aecd-fa7e9256a026)

How can I adjust the position of quantization nodes to reduce data conversion?

Through observing the automatic and manual placement of QDQ nodes in YOLOv7, I roughly understand that the parameters within the red box below need to be set equal. However, this...

How can I adjust the position of quantization nodes to reduce data conversion?

@Vieeo trt-engine-explorer

How can I release the GPU memory used by triton_python_backend_stub when using the Python backend?

I want to use python backend to server my onnx model, why don't use onnx backend? because It is too slow at that onnx version in practice, so I choose...

[Feature] 问一句，mask2former如何能不使用mmdeploy::grid_sampler而使用torch.grid_sampler

老铁，请教一下怎么解决的，需要修改哪里

[Feature] 问一句，mask2former如何能不使用mmdeploy::grid_sampler而使用torch.grid_sampler

我也需要先转onnx再转trt，咱们步骤还是有一些重合的，请问应该怎么修改，感激不尽