{"cells":[{"metadata":{},"cell_type":"markdown","source":"# Using Deep Learning to \"learn\" the images of earthquake audio data.\n\nI wanted to expand my knowledge of deep learning and specifically using fastai. This kernel explores how I converted the acoustic data from the LANL Earthquake competition into melspectogram images. After converting audio to images we can train an image regression model using the fastai library.\n\nSo far I haven't been able to get outstanding performance out of this approach, but I am new to deep learning and the fastai/pytorch library. Hopefully this will be helpful for others interested in learning more about deep learning."},{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"import numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nimport os\nimport matplotlib.pyplot as plt\nimport librosa.display\nfrom librosa.feature import melspectrogram\nimport librosa\nimport random\nfrom tqdm import tqdm_notebook\nimport random\nfrom tqdm import tqdm\nimport matplotlib.pyplot as plt\nimport matplotlib.image as mpimg","execution_count":1,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Loading the training labels and exploring the images"},{"metadata":{"trusted":true},"cell_type":"code","source":"# Train labels can be found in this CSV\ntrain_labels = pd.read_csv('../input/lanl-earthquake-spectrogram-images/training_labels.csv')\ntrain_labels.head()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_kg_hide-input":true},"cell_type":"code","source":"fig, (ax1, ax2, ax3, ax4, ax5) = plt.subplots(1, 5, figsize=(20, 10))\nimg1 = mpimg.imread('../input/lanl-earthquake-spectrogram-images/train_images_no_overlap/train_images_v3/train_0.png')\nimg2 = mpimg.imread('../input/lanl-earthquake-spectrogram-images/train_images_no_overlap/train_images_v3/train_100.png')\nimg3 = mpimg.imread('../input/lanl-earthquake-spectrogram-images/train_images_no_overlap/train_images_v3/train_200.png')\nimg4 = mpimg.imread('../input/lanl-earthquake-spectrogram-images/train_images_no_overlap/train_images_v3/train_300.png')\nimg5 = mpimg.imread('../input/lanl-earthquake-spectrogram-images/train_images_no_overlap/train_images_v3/train_400.png')\nax1.imshow(img1)\nax1.set_title('TTF - {:0.4f}'.format(train_labels.loc[train_labels['seg_id'] == 'train_0']['target'].values[0]), fontsize=25)\nax2.imshow(img2)\nax2.set_title('TTF - {:0.4f}'.format(train_labels.loc[train_labels['seg_id'] == 'train_100']['target'].values[0]), fontsize=25)\nax3.imshow(img3)\nax3.set_title('TTF - {:0.4f}'.format(train_labels.loc[train_labels['seg_id'] == 'train_200']['target'].values[0]), fontsize=25)\nax4.imshow(img4)\nax4.set_title('TTF - {:0.4f}'.format(train_labels.loc[train_labels['seg_id'] == 'train_300']['target'].values[0]), fontsize=25)\nax5.imshow(img5)\nax5.set_title('TTF - {:0.4f}'.format(train_labels.loc[train_labels['seg_id'] == 'train_400']['target'].values[0]), fontsize=25)\nax1.axis('off')\nax2.axis('off')\nax3.axis('off')\nax4.axis('off')\nax5.axis('off')\nplt.tight_layout()\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Example of how the images were created\n- We use the `librosa` package which converts audio into melspectograms.\n- I tweaked the sample rate `sr` number of mels `n_mels` and used a log scale `power_to_db`. The values I chose were mainly based off of what I thought made the images look unique and detailed. A more scientific approach to creating the melspectrograms may yield better results.\n"},{"metadata":{"trusted":true},"cell_type":"code","source":"seg_00030f = pd.read_csv('../input/LANL-Earthquake-Prediction/test/seg_00030f.csv')\ny = seg_00030f['acoustic_data'].astype('float').values","execution_count":35,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def plot_spectrogram(y, imgdir=None, imgname=None, plot=True, sr=10000, n_mels=1000, log_tf=True, vmin=-100, vmax=0):\n    # Let's make and display a mel-scaled power (energy-squared) spectrogram\n    #y = np.array([float(x) for x in df['acoustic_data'].values])\n    S = librosa.feature.melspectrogram(y, sr=sr, n_mels=n_mels)\n\n    # Convert to log scale (dB). We'll use the peak power (max) as reference.\n    if log_tf:\n        S = librosa.power_to_db(S, ref=np.max)\n    \n    if plot:\n        # Make a new figure\n        plt.figure(figsize=(15,5))\n        plt.imshow(S)\n        # draw a color bar\n        plt.colorbar(format='%+02.0f dB')\n        # Make the figure layout compact\n        plt.tight_layout()\n        plt.axis('off')\n    if imgname is not None:\n        plt.imsave('{}/{}.png'.format(imgdir, imgname), S)\n        plt.clf()\n        plt.close()\n    return","execution_count":36,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Plot an example using this function.\nplot_spectrogram(y)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Modeling with Fastai Library\nhttps://www.fast.ai/ is a great resource for those interested in learning more about deep learning. After watching a few of the fastai tutorial videos I felt comfortable creating a baseline model."},{"metadata":{"trusted":true},"cell_type":"code","source":"%reload_ext autoreload\n%autoreload 2\n%matplotlib inline\n\n# I personally hate that fastai encourages using `import *`\n# That is how it is taught in the course\nfrom fastai import *\nfrom fastai.vision import *","execution_count":39,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Create a fastai databunch of training images and labels.\n- You can use a random split, or provide train and valid indexes for determining the train/valid split.\n- I disabled transforms on the image but this may be useful when doing transfer learning.\n- label_cls must be set to FloatList as we are using the images to predict a float number (Time to failure)"},{"metadata":{"trusted":true},"cell_type":"code","source":"bs = 10\ntrain_labels['path'] = '../input/lanl-earthquake-spectrogram-images/train_images_no_overlap/train_images_v3/' + train_labels['seg_id'] + '.png'\n#valid_idx = train_labels[:10000].loc[train_labels['quake_number'].isin([1, 5, 8])].index\n#train_idx = train_labels[:10000].loc[~train_labels['quake_number'].isin([1, 5, 8])].index\ndata = (ImageList.from_df(train_labels[:-1], path='./', cols='path')\n        #.split_by_idxs(train_idx=train_idx, valid_idx=valid_idx)\n        .split_by_rand_pct(0.1)\n        .label_from_df('target', label_cls=FloatList)\n        #.transform(get_transforms(), size=255)\n        .databunch(bs=bs))","execution_count":42,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"## Checking the images in a databunch batch"},{"metadata":{"trusted":true},"cell_type":"code","source":"data.show_batch(rows=3, figsize=(7,6))","execution_count":43,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 504x432 with 9 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{},"cell_type":"markdown","source":"# Create the Fastai resnet50 model\n- Make sure you have GPU turned on or this will be very slow.\n- We are using mean_absolute_error as our evaluation metric"},{"metadata":{"trusted":true},"cell_type":"code","source":"def mean_absolute_error(pred:Tensor, targ:Tensor)->Rank0Tensor:\n    \"Mean absolute error between `pred` and `targ`.\"\n    pred,targ = flatten_check(pred,targ)\n    return torch.abs(targ - pred).mean()\n\nlearn = cnn_learner(data, models.resnet50, metrics=mean_absolute_error)\nlearn.fit_one_cycle(4, 0.01)","execution_count":46,"outputs":[{"output_type":"display_data","data":{"text/plain":"<IPython.core.display.HTML object>","text/html":"<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: left;\">\n      <th>epoch</th>\n      <th>train_loss</th>\n      <th>valid_loss</th>\n      <th>mean_absolute_error</th>\n      <th>time</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <td>0</td>\n      <td>10.691886</td>\n      <td>9.303072</td>\n      <td>2.388052</td>\n      <td>01:40</td>\n    </tr>\n    <tr>\n      <td>1</td>\n      <td>10.180142</td>\n      <td>11.678172</td>\n      <td>2.665515</td>\n      <td>01:39</td>\n    </tr>\n    <tr>\n      <td>2</td>\n      <td>8.405375</td>\n      <td>7.858608</td>\n      <td>2.270359</td>\n      <td>01:40</td>\n    </tr>\n    <tr>\n      <td>3</td>\n      <td>7.949627</td>\n      <td>6.747758</td>\n      <td>2.059102</td>\n      <td>01:39</td>\n    </tr>\n  </tbody>\n</table>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Plot train vs valid loss\nfig = learn.recorder.plot_losses(return_fig=True)\nfig.set_size_inches(15,5)","execution_count":76,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1080x360 with 1 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Unfreeze the model and search for a good learning rate\nlearn.unfreeze()\nlearn.lr_find()\nfig = learn.recorder.plot(return_fig=True)\nfig.set_size_inches(15,5)","execution_count":77,"outputs":[{"output_type":"display_data","data":{"text/plain":"<IPython.core.display.HTML object>","text/html":""},"metadata":{}},{"output_type":"stream","text":"LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n","name":"stdout"},{"output_type":"display_data","data":{"text/plain":"<Figure size 432x288 with 1 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"learn.fit_one_cycle(2, slice(1e-6, 3e-3/10))\nlearn.save('cnn-step1')","execution_count":78,"outputs":[{"output_type":"display_data","data":{"text/plain":"<IPython.core.display.HTML object>","text/html":"<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: left;\">\n      <th>epoch</th>\n      <th>train_loss</th>\n      <th>valid_loss</th>\n      <th>mean_absolute_error</th>\n      <th>time</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <td>0</td>\n      <td>7.592947</td>\n      <td>6.739468</td>\n      <td>2.054890</td>\n      <td>02:09</td>\n    </tr>\n    <tr>\n      <td>1</td>\n      <td>8.003979</td>\n      <td>6.931798</td>\n      <td>2.085333</td>\n      <td>02:09</td>\n    </tr>\n  </tbody>\n</table>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Export the model\nlearn.export()","execution_count":81,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"# We can see there is now an export.pkl file that we've saved\n!ls -l","execution_count":85,"outputs":[{"output_type":"stream","text":"total 100656\r\n-rw-r--r-- 1 root root    136875 May  8 17:30 __notebook_source__.ipynb\r\n-rw-r--r-- 1 root root 102817372 May  8 17:37 export.pkl\r\ndrwxr-xr-x 2 root root      4096 May  8 17:35 models\r\n","name":"stdout"}]},{"metadata":{},"cell_type":"markdown","source":"Validation MAE ~ 2.08 isn't that great and we are starting to overfit. Using a better cross validation technique may be necessary.\n\nI'm sure you can do better than I have here, but this should be enough to get you started."},{"metadata":{},"cell_type":"markdown","source":"# Predicting on the test set\n- We load our sample submission file and create an imagelist based off of the seg_id\n- We point this imagelist at the images we've creates for the test set.\n- Load the trained model and call the prediction method on this imagelist."},{"metadata":{"trusted":true},"cell_type":"code","source":"ss = pd.read_csv('../input/LANL-Earthquake-Prediction/sample_submission.csv')\ntest = ImageList.from_df(ss, '../input/lanl-earthquake-spectrogram-images/test_images/test_images_v3/', cols='seg_id', suffix='.png')\nlearn = load_learner('./', test=test)\nlearn.load('cnn-step1')\npreds = learn.get_preds(ds_type=DatasetType.Test)","execution_count":86,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Save the time to failure\nss['time_to_failure'] = [float(x) for x in preds[0]]","execution_count":87,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"ss.head()","execution_count":88,"outputs":[{"output_type":"execute_result","execution_count":88,"data":{"text/plain":"       seg_id  time_to_failure\n0  seg_00030f         4.873120\n1  seg_0012b5         4.994225\n2  seg_00184e         5.322777\n3  seg_003339         9.331327\n4  seg_0042cc         6.957149","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>seg_id</th>\n      <th>time_to_failure</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>seg_00030f</td>\n      <td>4.873120</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>seg_0012b5</td>\n      <td>4.994225</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>seg_00184e</td>\n      <td>5.322777</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>seg_003339</td>\n      <td>9.331327</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>seg_0042cc</td>\n      <td>6.957149</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Cap the minimum and maximum time to failure values\nss.loc[ss['time_to_failure'] < 0, 'time_to_failure'] = 0\nss.loc[ss['time_to_failure'] > 12, 'time_to_failure'] = 12","execution_count":98,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"ss.plot(kind='hist', bins=100, figsize=(15, 5), title='Distribution of predictions on the Test Set')\nplt.show()","execution_count":100,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1080x360 with 1 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Save our predictions\nss.to_csv('submission.csv', index=False)","execution_count":102,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"I hope this can be helpful to others. Please let me know if you have any suggestions on how I could improve it."},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.4","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}