Commit History
[GSoC] Gemm and MatMul block quantization support (#268) 0a88ce4
[GSoC] Blockwise Quantization Tool (#265) 6911e5d
bump opencv version to 4.10.0 (#260) 167c85e
Introduce Preprocessing for Optimized Quantization in `quantize-ort.py` (#238) 2129055
Bump version 4.9 (#222) 1047434
add text detection model from ppocrv3 (#180) ac5c83c
Wanli commited on
Remove DaSiamRPN since we have its superseder VitTrack now (#213) 7c978be
fix model quantization to support timvx backend (#209) 80eebd6
bump version to 4.8.0 (#190) 00a9b48
shorten int8-quantized naming (#149) 284d26a
replace pphumanseg with a simplified one (#143) b57baee
bump lpd_yunet opset version to 13 (#142) c62dd45
Update handpose estimation model from MediaPipe (2023feb) (#133) 50fc340
Wanli commited on
Update palm detection model from MediaPipe (2023feb) (#128) 260eb6d
Wanli commited on
Update FER quantized model and fix document (#129) ebeb80f
Wanli commited on
Add a model for facial expression recognition (#100) 596a24b
Chengrui Wang commited on
Re-quantify some models from per_channel mode to per_tensor mode (#90) 323da84
Wanli commited on
fix some error and bugs (#112) 90edc6d
Wanli commited on
update benchmark (#97) 8b5437e
Wanli commited on
revise the quantize-ort to quantize crnn into int8 with high accuracy (#94) d30c3db
Yiyao Wang commited on
Add hand pose estimation model from Mediapipe (#83) 57699b7
Wanli commited on
use QuantFormat.QOperator by default to avoid fake quantization (#88) 4b236af
Add the missing yaml config for quantizing MP-PalmDet and improve quantized MP-PalmDet (#60) 83c563e
Add LPD_YuNet for License Plate Detection (#56) 6eef315
add palm detector from mediapipe (#51) e5b568e
update mobilenets (#48) 86cd32d
add mobilenets and benchmarking results d33294b
fytao commited on
Use YuNet of fixed input shape to avoid 'parseShape' error (#45) 62917b7
Add tools for quantization and quantized models (#36) 55654c5
Yuantao Feng commited on