-
Python 2.7.6+, 3.5.1+
- Chainer 1.10.0+
- scikit-image 0.11.3
- scipy 0.16.0
- numpy 1.10. 1
$ bash download.sh
$ nohup python train.py &
$ python draw_loss.py --logfile log.txt --outfile log.jpg
$ python train.py --model models/ResNet.py --lr 0.1 --gpu 0
The test accuracy after 15 epochs is 0.9406 (error (%): 5.94). The test accuracy reported in the MSR paper (Deep Residual Learning for Image Recognition) is 0.9357 (error (%): 6.43) (see Table 6).