► 15. Функции активации 4 | Курс Нейронные Сети.

✅ Courses with problems: ► Neural Networks with problems: https://clck.ru/3QMA9q ► Pytorch with problems: https://clck.ru/3QKioT ► Numpy with problems: https://clck.ru/3QKipY ► Object Detection with problems: https://clck.ru/3QKiq5 ► Pandas with problems: https://clck.ru/3QKipC ✅ My Telegram channel: https://t.me/dubinin_ser ✅ Telegram groups: ► Pytorch: https://t.me/PyTorch_for_you ► Pandas: https://t.me/pandas_for_you ► Numpy: https://t.me/numpy_for_you Git - https://clck.ru/3QPBCH =================================================== This course will introduce you to the key concepts of neural networks, from simple perceptrons to deep learning methods. We'll cover the main architectures (MLP, CNN, and Transformers), training principles, optimization, and regularization. Each video will include clear visualizations and step-by-step explanations so you can quickly apply the knowledge you've gained. This course is suitable for developers, students, and anyone interested in solving problems in computer vision, text processing, and forecasting. By the end, you'll be able to build, train, and evaluate your own neural networks and integrate them into projects. ==================================================== In this video, we implement some activation functions and compare them. Timecodes: 00:00 - Introduction. 00:07 - ReLU implementation. 01:54 - ELU implementation. 03:00 - Swish implementation. 04:33 - Sigmoid implementation. 05:26 - Softmax implementation. 06:20 - Activation function comparison. 08:15 - Vanishing gradients. Tags: #pytorch #AI #objectdetection #neuralnet