How to Record the Entire ML-Agents Training Process as a Video?

Hi everyone,

I’m currently using Unity ML-Agents for training an AI agent, and I’d like to record the entire training process and output it as a video file. My goal is to capture everything happening in the game environment during training, even when the simulation is running at a high Time.timeScale (e.g., 10x or more).
Is there a way to do that?

To record the entire training process in Unity and save it as a video file you can use a few methods:

Unity Recorder: Unity has a built-in tool called Unity Recorder that allows you to record video, screenshots, and animations during gameplay. You can set it up to record the training process even with increased Time.timeScale. To do this:

  1. Go to Window > General > Recorder.
  2. Create a new recording task, select the recording type (e.g., video), and adjust the settings to record during the desired stage of the game.
  3. Start the recording and save the video file after it’s finished.
  4. Custom Recording with RenderTexture: If Unity Recorder doesn’t work for your needs, you can use RenderTexture to capture frames and create a video manually using third-party libraries (e.g., FFmpeg for saving frames as video). This gives you full control over the recording process.

Also, you can create an awesome cover for your video using tools from Creative Fabrica. They have lots of exciting new releases coming soon, and you can use their Font Generator and other design tools to make your video thumbnail look fantastic! Here’s the link to check it out: https://www.creativefabrica.com/the-artistry/creative-fabrica-news/a-new-era-of-generation-is-coming/