Skip to content

add rexnet-lite weights #132

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 3 commits into from
Feb 1, 2022
Merged

add rexnet-lite weights #132

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
0d5d669
add rexnet-lite weights
rentainhe Feb 1, 2022
c090565
update README
rentainhe Feb 1, 2022
f75039f
update changelog
rentainhe Feb 1, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,6 +55,7 @@ pip install flowvision==0.0.6
<li>ResMLP</li>
<li>gMLP</li>
<li>ConvMixer</li>
<li>ConvNeXt</li>
</ul>
<li><b>Detection</b></li>
<ul>
Expand Down
1 change: 1 addition & 0 deletions docs/source/changelog.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@
- Support `PolyLRScheduler` and `TanhLRScheduler` in `flowvision.scheduler` [#85](https://github.com/Oneflow-Inc/vision/pull/85)
- Add `resmlp_12_224_dino` model and pretrained weight [#128](https://github.com/Oneflow-Inc/vision/pull/128)
- Support [ConvNeXt](https://arxiv.org/abs/2201.03545) model [#93](https://github.com/Oneflow-Inc/vision/pull/93)
- Add [ReXNet](https://arxiv.org/abs/2007.00992) weights [#132](https://github.com/Oneflow-Inc/vision/pull/132)

**Bug Fixes**

Expand Down
2 changes: 1 addition & 1 deletion docs/source/tutorials/Getting Started.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@
## Installation
- To install latest stable release of flowvision:
```bash
pip install flowvision==0.0.56
pip install flowvision==0.0.6
```
- For an editable install
```bash
Expand Down
6 changes: 3 additions & 3 deletions flowvision/models/rexnet_lite.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,9 +11,9 @@

model_urls = {
"rexnet_lite_1_0": "https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/RexNet/rexnet_lite_1_0.zip",
"rexnet_lite_1_3": None,
"rexnet_lite_1_5": None,
"rexnet_lite_2_0": None,
"rexnet_lite_1_3": "https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/RexNet/rexnet_lite_1_3.zip",
"rexnet_lite_1_5": "https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/RexNet/rexnet_lite_1_5.zip",
"rexnet_lite_2_0": "https://oneflow-public.oss-cn-beijing.aliyuncs.com/model_zoo/flowvision/classification/RexNet/rexnet_lite_2_0.zip",
}


Expand Down
4 changes: 2 additions & 2 deletions projects/benchmark/classification/eval.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,14 +1,14 @@
export PYTHONPATH=$PWD:$PYTHONPATH
set -aux

MODEL="convnext_iso_large_224"
MODEL="rexnet_lite_1_5"
BATCH_SIZE=64
DATA_PATH="/dataset/imagenet/extract"
IMG_SIZE=224
NORMALIZE_MODE="IMAGENET_DEFAULT_MEAN_STD" # IMAGENET_INCEPTION_MEAN_STD or IMAGENET_DEFAULT_MEAN_STD or VIT_MIIL
CROP_PCT=0.875
INTERPOLATION="bicubic"
NUM_WORKERS=8
NUM_WORKERS=0
DEVICE=$1

CUDA_VISIBLE_DEVICES=$DEVICE python ./projects/benchmark/classification/benchmark.py --model $MODEL \
Expand Down
5 changes: 4 additions & 1 deletion results/results_imagenet.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,10 @@
| rexnetv1_1_5 | 80.314 | 19.686 | 95.162 | 4.838 | 9.7M | 224 | 0.875 | bicubic |
| rexnetv1_2_0 | 81.626 | 18.374 | 95.666 | 4.334 | 16.4M | 224 | 0.875 | bicubic |
| rexnetv1_3_0 | 82.792 | 17.208 | 96.242 | 3.758 | 34.7M | 224 | 0.875 | bicubic |
| rexnet_lite_1_0 | 76.176 | 23.824 | 92.824 | 7.176 | 4.8M | 224 | 0.875 | bicubic |
| rexnet_lite_1_0 | 76.176 | 23.824 | 92.824 | 7.176 | 4.8M | 224 | 0.875 | bicubic |
| rexnet_lite_1_3 | 77.726 | 22.274 | 93.788 | 6.212 | 6.7M | 224 | 0.875 | bicubic |
| rexnet_lite_1_5 | 78.588 | 21.412 | 94.192 | 5.828 | 8.3M | 224 | 0.875 | bicubic |
| rexnet_lite_2_0 | 80.134 | 19.866 | 95.006 | 4.994 | 13.0M | 224 | 0.875 | bicubic |
| vit_tiny_patch16_224 | 75.458 | 24.542 | 92.850 | 7.150 | 5.7M | 224 | 0.900 | bicubic |
| vit_tiny_patch16_384 | 78.440 | 21.560 | 94.546 | 5.454 | 5.7M | 384 | 1.000 | bicubic |
| vit_small_patch32_224 | 75.990 | 24.010 | 93.274 | 6.726 | 22.9M | 224 | 0.900 | bicubic |
Expand Down