From 89b0687cd1f61eac6c81bd6c908de0058003d298 Mon Sep 17 00:00:00 2001 From: Mingyu1991 <115005477+Mingyu1991@users.noreply.github.com> Date: Tue, 29 Nov 2022 19:22:13 +0900 Subject: [PATCH] docs: update traffic light ssd README.md (#2396) update traffic light ssd README.md --- .../traffic_light_ssd_fine_detector/README.md | 18 +++++++++++++----- 1 file changed, 13 insertions(+), 5 deletions(-) diff --git a/perception/traffic_light_ssd_fine_detector/README.md b/perception/traffic_light_ssd_fine_detector/README.md index 87ff47c84085e..3b283f35d2459 100644 --- a/perception/traffic_light_ssd_fine_detector/README.md +++ b/perception/traffic_light_ssd_fine_detector/README.md @@ -4,7 +4,19 @@ It is a package for traffic light detection using MobileNetV2 and SSDLite. -The trained model is based on [pytorch-ssd](https://github.com/qfgaohao/pytorch-ssd). +## Training Information + +### Pretrained Model + +The model is based on [pytorch-ssd](https://github.com/qfgaohao/pytorch-ssd) and the pretrained model could be downloaded from [here](https://storage.googleapis.com/models-hao/mb2-ssd-lite-mp-0_686.pth). + +### Training Data + +The model was fine-tuned on 1750 TierIV internal images of Japanese traffic lights. + +### Trained Onnx model + +- ## Inner-workings / Algorithms @@ -48,10 +60,6 @@ Based on the camera image and the global ROI array detected by `map_based_detect ## Assumptions / Known limits -## Onnx model - -- - ## Reference repositories pytorch-ssd github repository