Model training
Accelerate your model training and build something extraordinary! Leverage Nebius AI infrastructure powered by NVIDIA® H100 GPUs.
GPUs for different workloads
We offer AI-tailored NVIDIA A100 and H100 GPUs in DELTA HGX Baseboards with 8 GPUs connected by NVLink, as well as L40s and other GPUs in the PCIe form factor.
InfiniBand fabric
Get the most of multihost training on thousands of GPUs with full mesh connection. Our latest InfiniBand network supports up to 3.2Tb/s per host.
Marketplace
Leading vendors' AI-specific tools including OS images and Kubernetes® apps will make a perfect workplace for ML engineers.
How to choose GPU for training
V100
Good for dealing with small and middle-size models that do not require BF16 precision support.
Much more affordable than other types of GPU.
А100
Cost effective for single-node training of conventional models.
Great for domains where CNN, RNN models are popular, e. g. computer vision or medical diagnostics.
Н100
Best choice if speed is your top priority.
Perfect for bigNLP, LLM, and all models with Transformer architecture.
H200
The world’s most powerful GPU for supercharging AI and HPC workloads coming soon!
Solution architecture
Solution architecture
This set of Nebius AI services will enable you to create an environment and a data pipeline for supervised, semi-supervised, unsupervised or reinforcement learning.
Let’s find the best possible technical solution
If you want to use a specific database or third-party software for your project, our team of solution architects is here to assist you every step of the way.
FAQ & basic terminology
What is model training?
What is model training?
Model training is the process of teaching a machine learning algorithm to recognize patterns and make predictions or decisions based on data. During the training, the algorithm learns from labeled data, adjusting its parameters to minimize the difference between its predictions and the actual labels.
What is model retraining?
What is model retraining?
What are the benefits of using GPUs in the cloud for model training?
What are the benefits of using GPUs in the cloud for model training?
Is GPU acceleration necessary for all types of machine learning models?
Is GPU acceleration necessary for all types of machine learning models?
Can I use multiple GPUs for model training in the cloud?
Can I use multiple GPUs for model training in the cloud?