Language: Turkish
Organisers: TUBITAK ULAKBIM & Middle East Technical University
Format: Online
Level: Beginner
Prerequisites:
- Basic level experience in deep learning
- Familiarity with the Python programming language
- General understanding of deep learning concepts and principles
Topic: Data Science
General Information:
-
Introduction: How deep learning training works — what are backpropagation, loss functions, optimizers, and batches, and why they matter.
-
How does a single GPU work? Transferring data to the GPU; the sequence of forward pass, loss computation, backward pass, and optimizer steps; understanding what each line of code does.
-
What changes when moving to multi-GPU training? Data distribution across GPUs, gradient aggregation, and the additional synchronization overhead.
-
Live demo / code analysis: Monitoring data transfer, forward–backward passes, and parameter update timings; performance comparison between 1 GPU vs 2 GPUs.
Trainer: Assoc. Prof. Dr. Erdem Akagündüz is a faculty member at the Informatics Institute of Middle East Technical University (METU). His research focuses on computer vision, deep learning, pattern recognition, and image processing, and he has authored numerous academic publications and holds international patents in these fields.
Contact: ncc@ulakbim.gov.tr
- X: https://twitter.com/EuroCC_Turkey
- LinkedIN: https://www.linkedin.com/company/eurocc-turkey
- YouTube: https://www.youtube.com/c/EuroCCTurkey
