- Program:
- This tool is for TecArt algorithm team to quantize their custom neural network's pre-trained model.
- This tool can quantize PyTorch model to 8-bit fixed-point model for hardware acceleration.
- It's a general script to use.
- History:
- 2019/09/27 Albert Dong @ First release
- 2019/10/9 Albert Dong @ add get model params
- 2019/10/13 Albert Dong @ add muti_gpu support
- 2019/10/13 Albert Dong @ remove tqdm support
- 2019/10/13 Albert Dong @ re-add tqdm support
- 2019/10/13 Albert Dong @ print params
- 2019/10/21 Albert Dong @ fix BN bug
- 2019/10/22 Albert Dong @ take part code
- 2019/10/22 Albert Dong @ fix inout bugs
- License:
- GPL v3
- Bugs:
- fc inout quantization.
- Jobs:
- 断点续传
- 加速
- 新功能
- 开放api
-
Notifications
You must be signed in to change notification settings - Fork 0
This tool can quantize pytorch model to 8-bit fixed-point model for hardware acceleration.
License
zhiwei-dong/PyTorch-Quantize-Tool
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
About
This tool can quantize pytorch model to 8-bit fixed-point model for hardware acceleration.
Resources
License
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published