These measurements were performed as part of the 2018 Additive Manufacturing Benchmark Test Series (AM-Bench). This dataset and the associated experiments are part of a continuing series of controlled benchmark tests, in conjunction with a conference series, with two initial goals, 1) to allow modelers of Additive Manufacturing processes to test their simulations against rigorous, highly controlled additive manufacturing benchmark test data, and 2) to encourage additive manufacturing practitioners to develop novel mitigation strategies for challenging build scenarios. More information regarding the AMBench 2018 study can be found at www.nist.gov/ambench. For this year's challenge, numerous metal parts of the same geometry were created using an identical processing condition using a commercial powder bed fusion machine. The eight parts in total were manufactured in two builds. In situ thermal measurements of a select region on one of the parts within each build were acquired at 1800 frames per second. The part is a bridge structure geometry that has 12 legs of varying size (5 mm x 5 mm, 5 mm x 2.5 mm, and 0.5 mm x 5 mm), each leg is 5 mm tall, then uses a 45-degree overhang to transition into the bridge structure with a constant cross section. Each part is manufactured using 0.02 mm layer thickness, a programmed laser power of 195 W traveling at a scan speed of 800 mm/s, and the hatch spacing is 0.1 mm. The part is manufactured in 624 layers and the total build time nearly 9.5 hours. Details on the experiment can be found at www.nist.gov/ambench/amb2018-01-description, while related post-process measurement results can be found at www.nist.gov/ambench/benchmark-test-data.This dataset consists of thermal videos and MATLAB data structures for each layer. These are provided for each layer of the build and are grouped ten layers at a time in the provided zip files. The thermal videos provide an overview of the radiant temperature (not accounting for emissivity) measured during each layer, while the MATLAB structures contain the measurement data along with information on the camera timing, calibration, and process information. Two MATLAB functions are also provided. The first allows the measured radiant temperature to be converted into true temperature based on an assumed emissivity correction factor. The second function recreates the thermal video files. The second MATLAB function helps to provides context on how to interact with the MATLAB structures.For a detailed description of the dataset, please refer to the NIST Journal of Research publication, "Thermography of the Metal Bridge Structures Fabricated for the 2018 Additive Manufacturing Benchmark Test Series (AM-Bench 2018)." (in press)