-
Hi~ I have quantized YOLOv5s models with naive-ptq and advanced-ptq respectively. The mAP of basemodel is 36.3, while naive-ptq is 35.8 and advanced-ptq is 35.0. The quantized results of advanced-ptq …
-
The overall process can be divided into the following steps:
- [x] make each subgraph executable, this allows us to collect the inputs of each subgraph when inference with the Calibration data.
- …
-
-
### 🚀 The feature, motivation and pitch
@kimishpatel and I are planning to add int8 support to the quantized softmax operator (in D35621286), and we would like for the fx quantization workflow to be …
-
您好,我这边首先使用convert_bevdet_to_TRT.py这个脚本生成了tensorrt的engine文件,但是在测试tensorrt推理的时候发现其输出结果和pytorch的差异较大,请问这是什么原因呢?
-
I am trying to do evaluation on quantized model for yolov5. Followed the suggested steps and i am hitting this issue here. Looks like the cocoapi is not able to load the results file
root@xhdr7525dc…
-
### Issue Type
Documentation
### Source
pip (model-compression-toolkit)
### MCT Version
1.7.1
### OS Platform and Distribution
_No response_
### Python version
_No response_…
-
### 1. System information
- OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Colab connected to VM of GCP (high RAM is needed)
- TensorFlow installation (pip package or built from source):…
-
Have you guys tried pytorch_quantization on pointpillars or centerpoint? when I use pytorch_quantization on centerpoint and used QAT, the outcome is much worse than torch model, about 10% worse. PTQ i…
-
### 请提出你的问题
1.关于ERNIE-3.0模型的量化过程疑问:
![2022-11-16 16-52-11屏幕截图](https://user-images.githubusercontent.com/23071264/202133949-654fdba6-bda7-46eb-bfa0-9cfcf9d728e3.png)
![2022-11-16 16-53-33屏幕截图](http…