<div class="sc-kdrUpr eZtUed"><div class="sc-UEtKG dGqiYy sc-hDzlxo bEIZRR"><div class="sc-fqwslf gsqkEc"><div class="sc-cBQMlg kAHhUk"><h2 class="sc-dcKlJK sc-cVttbi gqEuPW ksnHgj">About Dataset</h2></div></div></div><div class="sc-fHzVOS cUYeeo"><div class="sc-davvxH nUNNB"><div style="min-height: 80px;"><div class="sc-etVRix jqYJaa sc-jCNfQM igJSrG"><h3>General</h3>

<p>Since imageNet weights are not ideal as a starting point for this task. I have retrained the model on the chest X dataset. You can use the models here as a starting point for your training. that will significantly boost your CV and LB. Unfortunately, I will not share the trained models for two reasons.<br>

1-    Keep the LB stable.<br>

2- A lot  of Kagglers are just copying the notebook and using the weights. "which was really disappointing" </p>

<h3>Notes</h3>

<p>To load the model you need to add a classifier (Linear Layer) because I didn't drop the last layers when I saved the weights.<br>

New starting points will be added soon.<br>

Results:<br>

single model, single fold ResNet200D CV:96.7, LB:96.7<br>

single model, single fold Inceptionv3 CV:95.7, LB:95.9<br>

single model, single fold DenseNet121 CV:94.9, LB:95.7</p>

<h3>Data</h3>

<p><a aria-label="https://www.kaggle.com/nih-chest-xrays/data (opens in a new tab)" target="_blank" href="https://www.kaggle.com/nih-chest-xrays/data">https://www.kaggle.com/nih-chest-xrays/data</a></p></div></div></div>