jo1990 / scnn-sad_pytorch

Pytorch implementation of "Learning Lightweight Lane Detection CNNs by Self Attention Distillation (ICCV 2019)"

SCNN-SAD Pytorch

Pytorch implementation of "Learning Lightweight Lane Detection CNNs by Self Attention Distillation (ICCV 2019)"

You can find the previous version here

Demo

Video

Demo trained with CULane dataset & tested with \driver_193_90frame\06051123_0635.MP4

gpu_runtime: 0.016253232955932617 FPS: 61

total_runtime: 0.017553091049194336 FPS: 56 on RTX 2080 TI

Comparison

Category	40k episode (before SAD)	60k episode (after SAD)
Image
Lane

Train

Requirements

• pytorch
• tensorflow (for tensorboard)

Datasets

• CULane

  CULane dataset path (click to expand)

  CULane_path
  ├─ driver_100_30frame
  ├─ driver_161_90frame
  ├─ driver_182_30frame
  ├─ driver_193_90frame
  ├─ driver_23_30frame
  ├─ driver_37_30frame
  ├─ laneseg_label_w16
  ├─ laneseg_label_w16_test
  └─ list
  

You need to change the correct dataset path in ./config.py

Dataset_Path = dict(
    CULane = "/workspace/CULANE_DATASET",
"
)

Training

First, change some hyperparameters in ./experiments/*/cfg.json

{
  "model": "enet_sad",               <- "scnn" or "scnn_sad" or "enet_sad"
  "dataset": {
    "dataset_name": "CULane",        <- "CULane" or "Tusimple"
    "batch_size": 12,
    "resize_shape": [800, 288]       <- [800, 288] with CULane, [640, 368] with Tusimple, and [640, 360] with BDD100K
                                        This size is defined in the ENet-SAD paper, any size is fine if it is a multiple of 8.
  },
  ...
}

And then, start training with train.py

python train.py --exp_dir ./experiments/exp1

Acknowledgement

This repo is built upon official implementation ENet-SAD and based on PyTorch-ENet, SCNN_Pytorch.