Netherlands F3 Interpretation Dataset Machine learning and, more specifically, deep learning algorithms have seen remarkable growth in their popularity and usefulness in the last years. Such a fact is arguably due to three main factors: powerful computers, new techniques to train deeper networks and more massive datasets. Although the first two are readily available in modern computers and ML libraries, the last one remains a challenge for many domains. It is a fact that big data is a reality in almost all fields today, and geosciences are not an exception. However, to achieve the success of general-purpose applications such as ImageNet - for which there are +14 million labeled images for 1000 target classes - we not only need more data, we need more high-quality labeled data. Such demand is even more difficult when it comes to the Oil & Gas industry, in which confidentiality and commercial interests often hinder the sharing of datasets to others. In this letter, we present the Netherlands interpretation dataset, a contribution to the development of machine learning in seismic interpretation. The Netherlands F3 dataset was acquired in the North Sea, offshore Netherlands. The data is publicly available and comprises pos-stack data, eight horizons and well logs of 4 wells. However, for the dataset to be of practical use for our tasks, we had to reinterpret the seismic, generating nine horizons separating different seismic facies intervals. The interpreted horizons were used to create 601 labeled masks for inlines and 482 for crosslines. We present the results of two experiments to demonstrate the utility of our dataset. Dataset contents Crosslines: Classes: 9 Number of slices: 482 Records per class: 864 Total of records: 7,776 Inlines: Classes: 9 Number of slices: 601 Records per class: 2,994 Total of records: 26,496 Configuration Crop: [0, 0, 0, 0] Gray levels: 256 Noise: 0.3 Percentile: 5.0 Strides: [20, 20] Tile shape: [40, 40, 1]