Skip to content

Commit

Permalink
readme
Browse files Browse the repository at this point in the history
  • Loading branch information
Richard Zhang authored and Richard Zhang committed Oct 22, 2020
1 parent 0c18b9a commit 010c3a8
Showing 1 changed file with 45 additions and 15 deletions.
60 changes: 45 additions & 15 deletions README_IMAGENET.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,19 +10,31 @@ We describe how to evaluate models for shift-invariance.

## (1) Evaluating models

We provide models with filter sizes 2,3,4,5 for AlexNet, VGG16, VGG16bn, ResNet18,34,50,101, DenseNet121, and MobileNetv2.

### Evaluating accuracy

```bash
python main.py --data /PTH/TO/ILSVRC2012 -a alexnet_lpf4 --pretrained --gpu 0 -e
python main.py --data /PTH/TO/ILSVRC2012 -a alexnet_lpf4 --pretrained -e --gpu 0
python main.py --data /PTH/TO/ILSVRC2012 -a vgg11_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a vgg13_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a vgg16_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a vgg19_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a vgg11_bn_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a vgg13_bn_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a vgg16_bn_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a vgg19_bn_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a resnet18_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a resnet34_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a resnet50_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a resnet101_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a resnet152_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a resnext50_32x4d --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a resnext101_32x8d --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a wide_resnet50_2_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a wide_resnet101_2_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a densenet121_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a densenet169_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a densenet201_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a densenet161_lpf4 --pretrained -e
python main.py --data /PTH/TO/ILSVRC2012 -a mobilenet_v2_lpf4 --pretrained -e
```

Expand All @@ -40,22 +52,38 @@ python main.py --data /PTH/TO/ILSVRC2012 -e -ens 5 -ens_sm -a resnet34_lpf4 --w
Same as above, but flag `-es` evaluates the shift-consistency -- how often two random `224x224` crops are classified the same.

```bash
python main.py --data /PTH/TO/ILSVRC2012 -a alexnet_lpf4 --pretrained --gpu 0 -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a vgg16_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a vgg16_bn_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnet18_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnet34_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnet50_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnet101_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a densenet121_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a mobilenet_v2_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a alexnet_lpf4 --pretrained -es -b 8 --gpu 0
python main.py --data /PTH/TO/ILSVRC2012 -a vgg11_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a vgg13_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a vgg16_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a vgg19_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a vgg11_bn_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a vgg13_bn_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a vgg16_bn_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a vgg19_bn_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnet18_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnet34_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnet50_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnet101_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnet152_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnext50_32x4d --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a resnext101_32x8d --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a wide_resnet50_2_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a wide_resnet101_2_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a densenet121_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a densenet169_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a densenet201_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a densenet161_lpf4 --pretrained -es -b 8
python main.py --data /PTH/TO/ILSVRC2012 -a mobilenet_v2_lpf4 --pretrained -es -b 8
```

Some notes:
- These line commands are very similar to the base PyTorch [repository](https://github.com/pytorch/examples/tree/master/imagenet). Change `_lpf#` with filter size (2,3,4,5).
- The example commands use our pretrained. You can them from your own training checkpoints by subsituting `--pretrained` for `--resume PTH/TO/CHECKPOINT`.

### Imagenet Results
### Filter size

In the paper, we experimented with filter size. For AlexNet, VGG16, VGG16bn, ResNet18,34,50,101, DenseNet121, and MobileNetv2, we provide models with filter sizes 2, 3, 4 (default), 5.

<img src='https://richzhang.github.io/antialiased-cnns/resources/imagenet_ind2_noalex_v2.jpg' align="right" width=380>

Expand Down Expand Up @@ -164,9 +192,10 @@ Antialiasing requires extra computation (but no extra parameters). Below, we mea

## (2) Training antialiased models

The following commands train antialiased AlexNet, VGG16, VGG16bn, ResNet18,34,50, and Densenet121 models with filter size 4. Best checkpoint will be saved `[[OUT_DIR]]/model_best.pth.tar`.
The following commands train antialiased AlexNet, VGG16, VGG16bn, ResNet18,34,50, and Densenet121 models with filter size 4 from scratch. Best checkpoint will be saved `[[OUT_DIR]]/model_best.pth.tar`.

```bash
# Training from scratch
python main.py --data /PTH/TO/ILSVRC2012 -a alexnet_lpf4 --out-dir alexnet_lpf4 --gpu 0 --lr .01
python main.py --data /PTH/TO/ILSVRC2012 -a vgg16_lpf4 --out-dir vgg16_lpf4 --lr .01 -b 128 -ba 2
python main.py --data /PTH/TO/ILSVRC2012 -a vgg16_bn_lpf4 --out-dir vgg16_bn_lpf4 --lr .05 -b 128 -ba 2
Expand All @@ -178,9 +207,10 @@ python main.py --data /PTH/TO/ILSVRC2012 -a densenet121_lpf4 --out-dir densenet1
python main.py --data /PTH/TO/ILSVRC2012 -a mobilenet_v2_lpf4 --out-dir mobilenet_v2_lpf4 --lr .05 --cos_lr --wd 4e-5 --ep 150
```

**New (Oct 2020) Fine-tuning antialiased models from baseline weights** The commands above train *from scratch*. You can fine-tune antialiased models starting from baseline model weights. I do this by turning on the `--finetune` flag, and performing the last 2/3 of training (by default, this means `lr` lowered by 10 times and `60 epochs`). This gets better results than training from scratch.
**New (Oct 2020)** The commands above train *from scratch*. You can now fine-tune antialiased models starting from baseline model weights. I do this by turning on the `--finetune` flag, and performing the last 2/3 of training; by default, this means `lr` lowered by 10 times and we use 60 `epochs`. This gets better results than training from scratch.

```bash
# Training from baseline weights
python main.py --data /PTH/TO/ILSVRC2012 -a resnet50_lpf4 --out-dir resnet50_lpf4 --lr .01 -ep 60 --finetune
```

Expand Down

0 comments on commit 010c3a8

Please sign in to comment.