Skip to content

Commit

Permalink
Rename LeNet notebooks, putting framework first
Browse files Browse the repository at this point in the history
This will make it easier to add PyTorch or JAX or other implementations and
clearly identify them in the directory structure and the README file.
  • Loading branch information
mbrukman committed Apr 3, 2024
1 parent 238b67e commit f60732e
Show file tree
Hide file tree
Showing 4 changed files with 23 additions and 23 deletions.
Original file line number Diff line number Diff line change
Expand Up @@ -35,9 +35,9 @@
"[colab-badge]: https://colab.research.google.com/assets/colab-badge.svg\n",
"[binder-badge]: https://static.mybinder.org/badge_logo.svg\n",
"\n",
"[github-keras-v1]: https://github.com/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_v1_basic_impl_in_Keras.ipynb\n",
"[colab-keras-v1]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_v1_basic_impl_in_Keras.ipynb\n",
"[binder-keras-v1]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_v1_basic_impl_in_Keras.ipynb"
"[github-keras-v1]: https://github.com/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_Keras_v1_basic_implementation.ipynb\n",
"[colab-keras-v1]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_Keras_v1_basic_implementation.ipynb\n",
"[binder-keras-v1]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_Keras_v1_basic_implementation.ipynb"
]
},
{
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -35,9 +35,9 @@
"[colab-badge]: https://colab.research.google.com/assets/colab-badge.svg\n",
"[binder-badge]: https://static.mybinder.org/badge_logo.svg\n",
"\n",
"[github-keras-v2]: https://github.com/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_v2_custom_Subsampling_layer_and_activation_in_Keras.ipynb\n",
"[colab-keras-v2]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_v2_custom_Subsampling_layer_and_activation_in_Keras.ipynb\n",
"[binder-keras-v2]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_v2_custom_Subsampling_layer_and_activation_in_Keras.ipynb"
"[github-keras-v2]: https://github.com/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_Keras_v2_custom_Subsampling_layer_and_activation.ipynb\n",
"[colab-keras-v2]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_Keras_v2_custom_Subsampling_layer_and_activation.ipynb\n",
"[binder-keras-v2]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_Keras_v2_custom_Subsampling_layer_and_activation.ipynb"
]
},
{
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -35,9 +35,9 @@
"[colab-badge]: https://colab.research.google.com/assets/colab-badge.svg\n",
"[binder-badge]: https://static.mybinder.org/badge_logo.svg\n",
"\n",
"[github-keras-v3]: https://github.com/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_v3_Subsamping_fixed_scaling_and_learning_rate_decay_in_Keras.ipynb\n",
"[colab-keras-v3]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_v3_Subsamping_fixed_scaling_and_learning_rate_decay_in_Keras.ipynb\n",
"[binder-keras-v3]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_v3_Subsamping_fixed_scaling_and_learning_rate_decay_in_Keras.ipynb"
"[github-keras-v3]: https://github.com/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_Keras_v3_Subsamping_fixed_scaling_and_learning_rate_decay.ipynb\n",
"[colab-keras-v3]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_Keras_v3_Subsamping_fixed_scaling_and_learning_rate_decay.ipynb\n",
"[binder-keras-v3]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_Keras_v3_Subsamping_fixed_scaling_and_learning_rate_decay.ipynb"
]
},
{
Expand Down
28 changes: 14 additions & 14 deletions lenet/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -10,11 +10,11 @@ as an extension of the custom Subsampling layer.

Available implementations:

| | Description | Input<br/>pixel<br/>range | Pooling<br/>layer | Activation | Learning<br/>rate | Library | Notebook |
|:---:| -------------- |:-------------------------:|:-----------------:|:----------:|:-----------------:|:-------:|:--------:|
| v1 | Basic impl | [0, 1] | MaxPool2D | tanh | 0.001 | Keras | [![View on GitHub][github-badge]][github-keras-v1] [![Open In Colab][colab-badge]][colab-keras-v1] [![Open in Binder][binder-badge]][binder-keras-v1] |
| v2 | Custom layer,<br/>activation | [0, 1] | Subsampling | scaled<br/>tanh | 0.001 | Keras | [![View on GitHub][github-badge]][github-keras-v2] [![Open In Colab][colab-badge]][colab-keras-v2] [![Open in Binder][binder-badge]][binder-keras-v2] |
| v3 | Custom layer,<br/>activation,<br/>scaling,<br/>learning rate | [-0.1, 1.175] | Subsampling | scaled<br/>tanh | schedule | Keras | [![View on GitHub][github-badge]][github-keras-v3] [![Open In Colab][colab-badge]][colab-keras-v3] [![Open in Binder][binder-badge]][binder-keras-v3] |
| Framework | Approach | Description | Input<br/>pixel<br/>range | Pooling<br/>layer | Activation | Learning<br/>rate | Notebook |
|:---------:|:--------:|:--------------:|:-------------------------:|:-----------------:|:----------:|:-----------------:|:--------:|
| Keras | v1 | Basic implementation | [0, 1] | MaxPool2D | tanh | 0.001 | [![View on GitHub][github-badge]][github-keras-v1] [![Open In Colab][colab-badge]][colab-keras-v1] [![Open in Binder][binder-badge]][binder-keras-v1] |
| Keras | v2 | Custom layer,<br/>activation | [0, 1] | Subsampling | scaled<br/>tanh | 0.001 | [![View on GitHub][github-badge]][github-keras-v2] [![Open In Colab][colab-badge]][colab-keras-v2] [![Open in Binder][binder-badge]][binder-keras-v2] |
| Keras | v3 | Custom layer,<br/>activation,<br/>scaling,<br/>learning rate | [-0.1, 1.175] | Subsampling | scaled<br/>tanh | schedule | [![View on GitHub][github-badge]][github-keras-v3] [![Open In Colab][colab-badge]][colab-keras-v3] [![Open in Binder][binder-badge]][binder-keras-v3] |

Our implementation is based on the following paper:

Expand All @@ -37,17 +37,17 @@ See also:
[colab-badge]: https://colab.research.google.com/assets/colab-badge.svg
[binder-badge]: https://static.mybinder.org/badge_logo.svg

[github-keras-v1]: LeNet_v1_basic_impl_in_Keras.ipynb
[colab-keras-v1]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_v1_basic_impl_in_Keras.ipynb
[binder-keras-v1]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_v1_basic_impl_in_Keras.ipynb
[github-keras-v1]: LeNet_Keras_v1_basic_implementation.ipynb
[colab-keras-v1]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_Keras_v1_basic_implementation.ipynb
[binder-keras-v1]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_Keras_v1_basic_implementation.ipynb

[github-keras-v2]: LeNet_v2_custom_Subsampling_layer_and_activation_in_Keras.ipynb
[colab-keras-v2]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_v2_custom_Subsampling_layer_and_activation_in_Keras.ipynb
[binder-keras-v2]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_v2_custom_Subsampling_layer_and_activation_in_Keras.ipynb
[github-keras-v2]: LeNet_Keras_v2_custom_Subsampling_layer_and_activation.ipynb
[colab-keras-v2]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_Keras_v2_custom_Subsampling_layer_and_activation.ipynb
[binder-keras-v2]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_Keras_v2_custom_Subsampling_layer_and_activation.ipynb

[github-keras-v3]: LeNet_v3_Subsamping_fixed_scaling_and_learning_rate_decay_in_Keras.ipynb
[colab-keras-v3]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_v3_Subsamping_fixed_scaling_and_learning_rate_decay_in_Keras.ipynb
[binder-keras-v3]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_v3_Subsamping_fixed_scaling_and_learning_rate_decay_in_Keras.ipynb
[github-keras-v3]: LeNet_Keras_v3_Subsamping_fixed_scaling_and_learning_rate_decay.ipynb
[colab-keras-v3]: https://colab.research.google.com/github/mbrukman/reimplementing-ml-papers/blob/main/lenet/LeNet_Keras_v3_Subsamping_fixed_scaling_and_learning_rate_decay.ipynb
[binder-keras-v3]: https://mybinder.org/v2/gh/mbrukman/reimplementing-ml-papers/main?filepath=lenet/LeNet_Keras_v3_Subsamping_fixed_scaling_and_learning_rate_decay.ipynb

[lenet-google-scholar]: https://scholar.google.com/citations?view_op=view_citation&hl=en&user=WLN3QrAAAAAJ&citation_for_view=WLN3QrAAAAAJ:u5HHmVD_uO8C
[lenet-wikipedia]: https://en.wikipedia.org/wiki/LeNet

0 comments on commit f60732e

Please sign in to comment.