Skip to content

update URLs to fix deprecated links. #29

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

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

update URLs to fix deprecated links. #29

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
6 changes: 3 additions & 3 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,14 +32,14 @@ AI systems by requiring less computation, fewer engineers, and less data,
to facilitate the giant market of edge AI and AIoT.

<p align="center">
<img src="https://hanlab.mit.edu/projects/tinyml/figures/background1.png" width="100%" />
<img src="https://hanlab18.mit.edu/projects/tinyml/figures/background1.png" width="100%" />
</p>
<p align="center">
<img src="https://hanlab.mit.edu/projects/tinyml/figures/background2.png" width="100%" />
<img src="https://hanlab18.mit.edu/projects/tinyml/figures/background2.png" width="100%" />
</p>

## Demo
[![Watch the video](https://hanlab.mit.edu/projects/tinyml/figures/mcunet_demo.png)](https://youtu.be/YvioBgtec4U)
[![Watch the video](https://hanlab18.mit.edu/projects/tinyml/figures/mcunet_demo.png)](https://youtu.be/YvioBgtec4U)

## Related Projects

Expand Down
12 changes: 6 additions & 6 deletions tinytl/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# TinyTL: Reduce Activations, Not Trainable Parameters for Efficient On-Device Learning [[website]](https://hanlab.mit.edu/projects/tinyml/tinyTL/)
# TinyTL: Reduce Activations, Not Trainable Parameters for Efficient On-Device Learning [[website]](https://hanlab18.mit.edu/projects/tinyml/tinyTL/)

```BibTex
@inproceedings{
Expand All @@ -13,27 +13,27 @@

## On-Device Learning, not Just Inference
<p align="center">
<img src="https://hanlab.mit.edu/projects/tinyml/tinyTL/figures/on-device-learning.png" width="80%" />
<img src="https://hanlab18.mit.edu/projects/tinyml/tinyTL/figures/on-device-learning.png" width="80%" />
</p>

## Activation is the Main Bottleneck, not Parameters
<p align="center">
<img src="https://hanlab.mit.edu/projects/tinyml/tinyTL/figures/acitvation-is-the-bottleneck.png" width="70%" />
<img src="https://hanlab18.mit.edu/projects/tinyml/tinyTL/figures/acitvation-is-the-bottleneck.png" width="70%" />
</p>

## Tiny Transfer Learning
<p align="center">
<img src="https://hanlab.mit.edu/projects/tinyml/tinyTL/figures/tinyTL.png" width="70%" />
<img src="https://hanlab18.mit.edu/projects/tinyml/tinyTL/figures/tinyTL.png" width="70%" />
</p>

## Transfer Learning Results
<p align="center">
<img src="https://hanlab.mit.edu/projects/tinyml/tinyTL/figures/tinytl_results_8.png" width="80%" />
<img src="https://hanlab18.mit.edu/projects/tinyml/tinyTL/figures/tinytl_results_8.png" width="80%" />
</p>

## Combining with Batch Size 1 Training
<p align="center">
<img src="https://hanlab.mit.edu/projects/tinyml/tinyTL/figures/tinytl_batch1.png" width="80%" />
<img src="https://hanlab18.mit.edu/projects/tinyml/tinyTL/figures/tinytl_batch1.png" width="80%" />
</p>

## Data Preparation
Expand Down
4 changes: 2 additions & 2 deletions tinytl/dataset_setup_scripts/make_aircraft.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,8 +64,8 @@ def main():
shutil.rmtree(os.path.join(dataset_path, 'fgvc-aircraft-2013b'))
os.remove(os.path.join(dataset_path, 'fgvc-aircraft-2013b.tar.gz'))

download_file('https://hanlab.mit.edu/tools/image_dataset_formats/aircraft/train.txt')
download_file('https://hanlab.mit.edu/tools/image_dataset_formats/aircraft/val.txt')
download_file('https://hanlab18.mit.edu/tools/image_dataset_formats/aircraft/train.txt')
download_file('https://hanlab18.mit.edu/tools/image_dataset_formats/aircraft/val.txt')

test_data()
train_data()
Expand Down
4 changes: 2 additions & 2 deletions tinytl/dataset_setup_scripts/make_food.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,8 +63,8 @@ def main():
shutil.rmtree(os.path.join(dataset_path, 'food-101'))
os.remove(os.path.join(dataset_path, 'food-101.tar.gz'))

download_file('https://hanlab.mit.edu/tools/image_dataset_formats/food_101/train.txt')
download_file('https://hanlab.mit.edu/tools/image_dataset_formats/food_101/val.txt')
download_file('https://hanlab18.mit.edu/tools/image_dataset_formats/food_101/train.txt')
download_file('https://hanlab18.mit.edu/tools/image_dataset_formats/food_101/val.txt')

test_data()
train_data()
Expand Down
4 changes: 2 additions & 2 deletions tinytl/tinytl_fgvc_train.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -135,7 +135,7 @@
# replace bn layers with gn layers
replace_bn_with_gn(net, gn_channel_per_group=8)
# load pretrained model
init_file = download_url('https://hanlab.mit.edu/projects/tinyml/tinyTL/files/'
init_file = download_url('https://hanlab18.mit.edu/projects/tinyml/tinyTL/files/'
'proxylessnas_mobile+lite_residual@imagenet@ws+gn', model_dir='~/.tinytl/')
net.load_state_dict(torch.load(init_file, map_location='cpu')['state_dict'])
net.classifier = LinearLayer(
Expand All @@ -147,7 +147,7 @@
net_config_path = os.path.join(args.net_path, 'net.config')
init_path = os.path.join(args.net_path, 'init')
else:
base_url = 'https://hanlab.mit.edu/projects/tinyml/tinyTL/files/specialized/%s/' % args.dataset
base_url = 'https://hanlab18.mit.edu/projects/tinyml/tinyTL/files/specialized/%s/' % args.dataset
net_config_path = download_url(base_url + 'net.config',
model_dir='~/.tinytl/specialized/%s' % args.dataset)
init_path = download_url(base_url + 'init', model_dir='~/.tinytl/specialized/%s' % args.dataset)
Expand Down