Learn how to create a video classification model using Keras and TensorFlow. Instead of using 2D convolutions, we’ll be discussing how to use 3D convolutions. After we build that understanding, we’ll delve deeper into creating a more complex video classification model, known as a (2+1)D convolutional neural network.
Since the code is already available in the video classification tutorial linked below, we’ll focus more on the concepts in this video.We hope this video proves useful to you and allows you to build custom, sophisticated models for your own video classification tasks.
Resources:
You can find the code from this video here → https://goo.gle/3y6m4zx
Loading and preprocessing video data tutorial → https://goo.gle/3Xxc10G
Video classification tutorial with complete code → https://goo.gle/3Zao8SL
Transfer learning with MoViNet → https://goo.gle/40ZtDFu
Making new Layers and Models via subclassing → https://goo.gle/3Jh2qY9
Watch more ML Tech Talks → https://goo.gle/ml-tech-talks
Watch more ML Tech Talks → https://goo.gle/ml-tech-talks
Subscribe to TensorFlow → https://goo.gle/TensorFlow