The new instructor-led workshops deal with fundamentals of deep discovering, recommender systems and Transformer-based purposes. Anyone connected on line can be a part of for a nominal charge, and contributors will have obtain to a completely configured, GPU-accelerated server in the cloud.
DLI teacher-led trainings consist of fingers-on remote studying taught by NVIDIA-qualified specialists in virtual classrooms. Individuals can interact with their instructors and friends in real time. They can whiteboard concepts, tackle interactive coding difficulties and earn a DLI certificate of issue competency to aid their skilled expansion.
New DLI workshops launching at GTC include things like:
- Fundamentals of Deep Learning — Construct the self esteem to choose on a deep understanding project by learning how to coach a product, work with widespread knowledge types and model architectures, use transfer discovering concerning products, and extra.
- Creating Smart Recommender Systems — Make unique styles of recommender devices: written content-primarily based, collaborative filtering, hybrid, and additional. Understand how to use the open-source cuDF library, Apache Arrow, alternating the very least squares, CuPy and TensorFlow 2 to do so.
- Constructing Transformer-Primarily based Pure Language Processing Applications — Learn about NLP subjects like Word2Vec and recurrent neural community-dependent embeddings, as perfectly as Transformer architecture features and how to make improvements to them. Use pre-properly trained NLP products for text classification, named-entity recognition and query answering, and deploy refined models for stay applications.
Other DLI choices at GTC will involve:
- Fundamentals of Accelerated Computing with CUDA Python — Dive into how to use Numba to compile NVIDIA CUDA kernels from NumPy common capabilities, as nicely as make and launch customized CUDA kernels, although making use of essential GPU memory administration approaches.
- Programs of AI for Predictive Upkeep — Leverage predictive upkeep and detect anomalies to deal with failures and stay clear of high-priced unplanned downtimes, use time-collection facts to forecast results applying machine mastering classification styles with XGBoost, and extra.
- Fundamentals of Accelerated Facts Science with RAPIDS — Learn how to use cuDF and Dask to ingest and manipulate large datasets straight on the GPU, implementing GPU-accelerated equipment finding out algorithms which includes XGBoost, cuGRAPH and cuML to carry out data assessment at substantial scale.
- Fundamentals of Accelerated Computing with CUDA C/C — Locate out how to accelerate CPU-only apps to run their latent parallelism on GPUs, utilizing methods like vital CUDA memory management to improve accelerated programs.
- Fundamentals of Deep Understanding for Multi-GPUs — Scale deep learning coaching to a number of GPUs, appreciably shortening the time expected to train lots of data and earning resolving advanced challenges with deep studying feasible.
- Programs of AI for Anomaly Detection — Explore how to put into practice numerous AI-based mostly alternatives to determine community intrusions, working with accelerated XGBoost, deep understanding-primarily based autoencoders and generative adversarial networks.
With far more than two million registered NVIDIA builders performing on technological breakthroughs to clear up the world’s toughest challenges, the desire for deep finding out knowledge is larger than at any time. The complete DLI system catalog involves a assortment of subjects for everyone interested in mastering much more about AI, accelerated computing and details science.
Get a glimpse of the DLI encounter:
Workshops have confined seating, with the early chook deadline on Sep 25. Sign up now.