Best practices for GPU-accelerated instances

Deploy a distilled DeepSeek-R1 model on a GPU-accelerated instance

Deploy an inference service for a distilled DeepSeek-R1 model on a GPU-accelerated instance for lightweight, high-performance inference.

Deploy a full DeepSeek model in a distributed two-machine setup on GPU-accelerated instances

Use vLLM on two ebmgn8v instances to build a distributed inference service for DeepSeek-V3/R1, scaling computing power for large model training or high-concurrency inference.

Deploy a full DeepSeek model on a single GPU-accelerated instance

Use SGLang on a single ebmgn8v instance to build an inference service for DeepSeek-V3/R1 in a multi-GPU setup.

Build a model inference environment using a DeepGPU-LLM image

Build an LLM inference environment on a GPU-accelerated instance with the Alibaba Cloud DeepGPU-LLM container image, suited for NLP scenarios such as chatbots, text analytics, and coding assistants.

Quickly build a model inference environment using a vLLM image

Build an LLM inference environment on a GPU-accelerated instance with a vLLM container image, suited for NLP scenarios such as chatbots and text classification.

Accelerate text-to-image generation using an SD-WebUI container image

Deploy an SD-WebUI container image on a GPU-accelerated instance to accelerate text-to-image generation.

Deploy an NGC environment to build a deep learning development environment

Deploy an NGC environment on a GPU-accelerated instance to set up a deep learning development environment, using TensorFlow as an example.

Use FastGPU to one-click deploy and train applications

Follow hands-on FastGPU training tutorials in the Alibaba Cloud Developer Lab with a real environment.

Quickly configure eRDMA using an eRDMA image

Configure eRDMA on a GPU-accelerated instance with an eRDMA image to enable high-performance network communication in container environments such as Docker.

Build a model inference environment using TensorRT-LLM

Install and use TensorRT-LLM on a GPU-accelerated instance to build an LLM inference environment, suited for NLP scenarios such as chatbots and text analytics.

Build a secure large language model inference environment on heterogeneous confidential computing instances

Build a secure LLM inference environment with Alibaba Cloud heterogeneous confidential computing instances and the Confidential AI solution to protect data during inference.