 Hi, I'm William. Now, in GitLab 13.9, you can configure your GitLab runners to leverage GPUs in the Docker executor. Why is this great? It means that for those of you running computational workloads that can benefit from having a GPU, such as machine learning pipelines, now you can configure a dedicated runner to execute those jobs using a GPU. Let's consider this example. Here I have my project where I have been working on experiments and visualizing data. Now I want to retrain the model, changing some epochs in the main script. For that, I can use the Web IDE and let's change the epochs from 10 to 15. As you can see here, I'm using TensorFlow. Let's commit the changes. Add some commit message. Meaningful one. All right, now that I have committed the changes, let's take a look at the pipeline. As you can see, we have different stages from validation, training, testing and reporting. Only during the job called train model is when we are using our runner enabled with GPU. Let's take a look at that job. As you can see here, the job is starting and we are using GitLab Runner 13.9 with Docker executor. In the log, now we can see how the epochs start to execute. Now we know that we are using the GPU, as you can see in the text that we have created here, where we are using a dedicated runner with GPU enabled for this job. Okay, it's done. It finished training the model. Now if we check the next job, the reporting one, we can see that it is using one of the shared runners. We don't have the GPU tag because for this job, it is not necessary to have a GPU enabled machine. Okay, that was it. More information about how to enable GitLab runners with GPUs in the video description. Stay tuned and let's continue learning at GitLab.