-
### Checklist
- [X] I have searched related issues but cannot get the expected help.
- [X] 2. I have read the [FAQ documentation](https://github.com/open-mmlab/mmdeploy/tree/main/docs/en/faq.md) b…
-
## Description
## Environment
**TensorRT Version**: 8.5.1
**NVIDIA GPU**: A100
I used `https://github.com/NVIDIA/TensorRT/blob/main/demo/BERT/builder.py` to build int8 PTQ tinybert. Bu…
-
I want resnet50.wts quantization to fp32, fp16, int8 at resnet50.py. How do I modify resnet50.py to do each?
-
```
#!/usr/bin/python3
import argparse
parser = argparse.ArgumentParser()
parser.add_argument("input",type=str)
parser.add_argument("--output",type=str,default="display://0")
parse…
-
## Description
I've noticed that TensorRT inference speed, in general, is significantly slower when using `LayerNormalization` instead of `BatchNormalization`. Also in particular, any gains I've se…
-
## Description
trtexec --onnx=model.onnx --best(or --int8)
[05/11/2023-18:20:45] [E] Error[2]: [weightConvertors.cpp::computeStaticKernelScales::440] Error Code 2: Internal Error (Assertion …
-
## ❓ Question
I have been trying to use INT8 inference for a trained pytorch model.
I followed this:
https://pytorch.org/TensorRT/_notebooks/vgg-qat.html
and
https://docs.nvidia.com/deepl…
-
## Bug Description
When I try to run example/fx/quantized_resnet_test.py, it gives AssertionError which assert engine failed. The error occurs because it failed to find scale for int8 mode.
## …
-
I convert the model to TensorRT by this command
python tools/convert_bevdet_to_TRT.py configs/bevdet/bevdet-r50-cbgs.py checkpoint/bevdet-r50-cbgs.pth work_dir --fuse-conv-bn --fp16 --int8
Traceba…
-
I'm new to coding. During the running test, there are some objects is detected. I want to show the detected results in the video like counting number of people and what should I do? I don't know what …