cleanlab
diff --git a/‎.gitignore
+6 b/‎.gitignore
+6
diff --git a/‎README.md
+46-13 b/‎README.md
+46-13
diff --git a/‎amazon-reviews-fasttext/README.md
+17 b/‎amazon-reviews-fasttext/README.md
+17
@@ -0,0 +1,6 @@
+env/
+temp/
+output/
+data/
+.ipynb_checkpoints/
+cifar10-cnn-coteaching/data/
@@ -1,21 +1,54 @@
-# ``cleanlab`` Examples
+# cleanlab Examples
 
-Not sure where to start? Try checking out how to find [ImageNet Label Errors](imagenet/imagenet_train_label_errors.ipynb).
+This repo contains code examples that demonstrate how to use [cleanlab](https://github.com/cleanlab/cleanlab) and how [confident learning](https://arxiv.org/abs/1911.00068) works to find label errors.
 
+To quickly learn the basics of running cleanlab on your own data, we recommend first starting [here](https://docs.cleanlab.ai/) before diving into the examples below.
 
-A brief description of the files and folders:
-* `imagenet`, 'cifar10', 'mnist' - code to find label errors in these datasets and reproduce the results in the [confident learning paper](https://arxiv.org/abs/1911.00068). You will also need to `git clone` [confidentlearning-reproduce](https://github.com/cgnorthcutt/confidentlearning-reproduce).
-  - [imagenet_train_crossval.py](imagenet/imagenet_train_crossval.py) - a powerful script to train cross-validated predictions on ImageNet, combine cv folds, train with on masked input (train without label errors), etc.
-  - [cifar10_train_crossval.py](cifar10/cifar10_train_crossval.py) - same as above, but for CIFAR.
-* `classifier_comparison.ipynb` - tutorial showing `cleanlab` performance across 10 classifiers and 4 dataset distributions.
-* `iris_simple_example.ipynb` - tutorial showing how to use `cleanlab` on the simple IRIS dataset.
-* `model_selection_demo.ipynb` - tutorial showing model selection on the cleanlab's parameter settings.
-* `simplifying_confident_learning_tutorial.ipynb` - tutorial implementing cleanlab as raw numpy code.
-* `visualizing_confident_learning.ipynb` - tutorial to demonstrate the noise matrix estimation performed by cleanlab.
+## Table of Contents
+
+Recommended order of examples to try:
+
+|     | Example                                                                                        | Description                                                                                                                                                                                                                                                                  |
+| --- | ---------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| 1   | [iris_simple_example.ipynb](iris_simple_example.ipynb)                                         | Use cleanlab to find synthetic label errors in the Iris dataset.                                                                                                                                                                                                             |
+| 2   | [classifier_comparison.ipynb](classifier_comparison.ipynb)                                     | Demonstrate how cleanlab can be used to train 10 different classifiers on 4 dataset distributions with label errors.                                                                                                                                                         |
+| 3   | [model_selection_demo.ipynb](model_selection_demo.ipynb)                                       | Perform hyperparameter optimization to find the best settings of cleanlab's optional parameters.                                                                                                                                                                             |
+| 4   | [simplifying_confident_learning_tutorial.ipynb](simplifying_confident_learning_tutorial.ipynb) | Implement cleanlab as raw numpy code.                                                                                                                                                                                                                                        |
+| 5   | [visualizing_confident_learning.ipynb](visualizing_confident_learning.ipynb)                   | Demonstrate how cleanlab performs noise matrix estimation.                                                                                                                                                                                                                   |
+| 6   | [cifar10-cnn-coteaching](cifar10-cnn-coteaching)                                               | Demonstrate the use of two experimental modules from cleanlab: [cifar_cnn.py](https://github.com/cleanlab/cleanlab/blob/master/cleanlab/experimental/cifar_cnn.py) and [coteaching.py](https://github.com/cleanlab/cleanlab/blob/master/cleanlab/experimental/coteaching.py) |
+| 7   | [mnist-cnn](mnist-cnn)                                                                         | Demonstrate the use of the following experimental module from cleanlab: [mnist_pytorch.py](https://github.com/cleanlab/cleanlab/blob/master/cleanlab/experimental/mnist_pytorch.py)                                                                                          |
+| 8   | [amazon-reviews-fasttext](amazon-reviews-fasttext)                                             | Demonstrate the use of the following experimental module from cleanlab: [fasttext.py](https://github.com/cleanlab/cleanlab/blob/master/cleanlab/experimental/fasttext.py)                                                                                                    |
+
+## Instructions
+
+To run the latest example notebooks, execute the commands below which will install the required libraries in a virtual environment.
+
+It is recommended to run the examples with the latest stable cleanlab release. See `requirements.txt` file.
+
+```console
+$ python -m pip install virtualenv
+$ python -m venv env
+$ source env/bin/activate
+$ python -m pip install -r requirements.txt
+```
+
+For examples 1-5, you may run the notebooks individually or run the bash script below which will execute and save each notebook.
+
+Bash script:
+
+```console
+$ bash ./run_all_notebooks.sh
+```
+
+For examples 6-8, please follow the instructions in the `README` of each folder.
+
+## Older Examples
+
+See the `contrib` folder for examples from v1 of cleanlab which may be helpful for reproducing results from the [Confident Learning paper](https://arxiv.org/abs/1911.00068).
 
 ## License
 
-Copyright (c) 2017-2021 Cleanlab Inc.
+Copyright (c) 2017-2022 Cleanlab Inc.
 
 All files listed above and contained in this folder (<https://github.com/cleanlab/examples>) are part of cleanlab.
 
@@ -26,7 +59,7 @@ the Free Software Foundation, either version 3 of the License, or
 
 cleanlab is distributed in the hope that it will be useful,
 but WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
 GNU Affero General Public License for more details.
 
 You should have received a copy of the GNU Affero General Public License in [LICENSE](LICENSE).
@@ -0,0 +1,17 @@
+# Training a Fasttext model on the amazon reviews dataset
+
+This example demonstrates the use of the following experimental module below from cleanlab:
+
+- [cleanlab.experimental.fasttext.py](https://github.com/cleanlab/cleanlab/blob/master/cleanlab/experimental/fasttext.py)
+
+The code is adapted from cleanlab v1 examples (see `contrib/v1` folder).
+
+## Instructions
+
+Run bash script below to download all the data.
+
+```console
+$ bash ./download_data.sh
+```
+
+Start Jupyter Lab and run the notebook: `amazon_pyx.ipynb`