A TensorFlow implementation of ResNet-18(https://arxiv.org/abs/1512.03385)
- TensorFlow 1.8
- The ImageNet dataset
- All image files are required to be valid JPEG files. See this gist.
- It is highly recommened for every image to be resized so that the shorter side is 256.
- (Optional) Torchfile(to convert ResNet-18 .t7 checkpoint into tensorflow checkpoint. Install with a command
pip install torchfile)
How To Run
- (Optional) Convert torch .t7 into tensorflow ckpt
# Download the ResNet-18 torch checkpoint wget https://d2j0dndfm35trm.cloudfront.net/resnet-18.t7 # Convert into tensorflow checkpoint python extract_torch_t7.py
train_scratch.sh(training from scratch) or
train.sh(finetune pretrained weights) to have valid values of following arguments
val_image_root: Path to the list file of train/val dataset and to the root
num_gpusand corresponding IDs of GPUs(
CUDA_VISIBLE_DEVICESat the first line)
./train.shif you want to finetune the converted ResNet(NOTE: The model needs to be finetuned for some epochs)
./train_scratch.shif you want to train ResNet from scratch
- Evaluate the trained model
./eval.shfor evaluating the trained model(change the arguments in
eval.shto your preference)
- The extracted weights should be finetuned for several epochs(run
./train.sh) to get the full performance(If you run the evaluation code without finetuning, the single-crop top-1 validation accuracy is about 60%, which is less than the appeared in the original). I guess there is some minor issue that I have missed.