deepglint / deepglint/EasyQuant
FaceDetection Task bad result?
Open
- Dominant language
- Python
- Stars
- 407
- Forks
- 67
- PR merge metrics
- No merged PRs in 30d
Description
Origin question is here https://github.com/deepglint/EasyQuant/issues/3#issuecomment-653895074 ;
More clear:
1. ncnn origin is ncnn origin post-training quantization TRT KLD function result, it almost same as origin float32 result.
2. I will check my activation distribution
3. I notice some layer get bad cosine distance val, you can see the below log
[EQ.log](https://github.com/deepglint/EasyQuant/files/4879925/EQ.log)
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.