-
I am tring to obtain a QAT model with official [pytorch-quantization](https://github.com/NVIDIA/TensorRT/tree/master/tools/pytorch-quantization).
However, I notice that
(1) the quantization func…
-
您好，非常感谢你们开源的优秀项目。最近想学习一下，我看了下文档，好像没找到感知训练量化的内容(可能是自己粗心没找到)。如果是我没找到希望您能稍微指点一下文档在哪，如果真的没有，想了解下您们的一些想法。谢谢。
-
## Description
After I used onnx-tensorrt to complete the int8 quantization of the resnet18 model, I found that the performance was the same as that of fp16 (batchsize=64). I would like to ask if t…
-
**System information**
- TensorFlow version (you are using): TF 2.3.0
- Are you willing to contribute it (Yes/No): Yes
**Motivation**
1. I think there are many users who want to accelerate…
-
-
From QDROP paper，i notice the benchmark result include Faster RCNN；
Could you provid…
-
# Describe the bug
Use `nncase1.7.1` to convert the `shufflenet_v2_x0_5` model in TorchVision, the quantization is stuck in 4.3. Quantize graph... , and ctrl+c does not respond. If quantization …
-
!!!!!!!!!!首先声明!!!!!!!!!!
这个问题是基于commit:a5582f416d62d9c40d3f5023b81ce71bb8791dd9 进行的，因为当前的master版本导出tengine的时候scale文件是空的，所以基于这个版本进行转换。
使用这个版本的mqbench对resnet进行qat训练，qat训练精度基本保持不变，acc在90左右，然后使用conver…
-
### Description
```shell
Branch: up to date version
GPU: 8 x Tesla V100 32G
```
### Reproduced Steps
```shell
1. Follow the steps of docs/vit_guide.md
2. ......
3. cd examples/pytorch/vit/ViT-q…
-
https://github.com/PaddlePaddle/PaddleDetection/issues/5719#issue-1206059590
延续上面的问题
换了ppdet2.2/2.3，推理过程分别使用了inference benchmark官方github提供量化导出模型和自己训练的模型，fp16和fp32都可以正常推理，只有int8不行。
训练过程中确定使用了s…