Release Notes and Announcements
- Image Release Notes
Product Introduction
- Overview
- Strengths
- Scenarios
- Editions and Support
Compatibility Notes
- Hardware Compatibility List
- Commercial Software Compatibility List
System Update
- Patch Installation Guide
- Security Remediation Recommendation Commands: Precisely Executing Critical Security Updates
User Guide
- Kernel Features
- System Configuration
- Security Configuration
- Network Configuration
- Storage and File Systems
- Development Guide
- Performance Tuning
- System Diagnosis and Observability
TencentOS Scenario Images
- TencentOS AI Enhanced Image
- TencentOS AI Container Image
Practice Tutorial
- Docker Container User Guide
- NVIDIA Deployment Practices
FAQs
- Pre-Purchase FAQs
- Troubleshooting DNF/YUM Issues
Contact Us
Glossary

TencentOS AI Container Image

Baixar

Modo Foco

Tamanho da Fonte

Última atualização: 2026-10-19 15:10:53

Traduzido por IA

TencentOS AI container images are an out-of-the-box AI container image set officially optimized and maintained by Tencent Cloud. They come pre-installed with adapted and optimized AI frameworks, acceleration libraries, and runtime environments, providing a seamless AI development and deployment experience for various user scenarios. In addition, we perform compatibility adaptation for mainstream GPU models worldwide. TencentOS AI container images can help you build a stable, high-performance AI application environment while reducing the complexity and cost of environment configuration.
﻿
Types of TencentOS AI Container Images
TencentOS AI container images provide you with three types of core images: BaseOS container images, framework container images, and large model container images. The image definitions and typical application scenarios are as follows:
Image type
Image Definition
Typical Scenarios
BaseOS Container Image
Provides a stable and lightweight TencentOS base container environment, making it easy for users to build custom images.
Suitable for users to build customized AI images or serve as the underlying foundation for their own applications.
Framework Container Image
Comes with nearly twenty mainstream AI frameworks preinstalled on top of the base image, covering various scenarios such as training and inference.
Suitable for users to directly perform model training and inference based on it.
LLM Container Image
Further integrates a complete toolchain and large model applications on top of the framework image, supporting out-of-the-box model services.
Suitable for users to directly deploy services and shorten the service launch cycle.
 
Features and Strengths
Environment consistency: All images have undergone compatibility testing and optimization, resolving dependency conflicts and environment configuration challenges.
Out-of-the-box: covers environments such as AI frameworks, AI models, acceleration libraries, and runtimes, enabling users to deploy and use them with one click.
Ops method: Based on different system images, users can use Docker Pull or yum to directly run them or perform system Ops management.
Enterprise-grade service: All images are rigorously tested by Tencent and receive long-term stable version updates, ensuring software supply chain security.
Ecosystem support: While inheriting the base system ecosystem, the epol repository from the OpenCloudOS community is introduced, allowing users to install various AI software using yum.
﻿
BaseOS Container Image
Image Name
Image address
Image description
TencentOS 4 minimal container image
docker.io/tencentos/tencentos4-minimal
The TencentOS 4 minimal container image is a minimal base image with full software package management features, using dnf as the package manager and coreutils as the Linux toolset.
﻿
Framework Image 
Image Name
CPU/GPU
Version information
Image address
Image description
tencentos4-pytorch
GPU
PyTorch: 2.8.0
CUDA: 12.9.1
NCCL: 2.27.6
cuDNN 9: 9.11.0.98
cuDNN 8: 8.9.7.29
Python: 3.11.6
OS: TencentOS 4
docker.io/tencentos/tencentos4-pytorch
A PyTorch 2.8.0 GPU Docker image based on TencentOS4 and CUDA 12.9. The image includes the official PyTorch example code repository for quick environment verification and learning.
tencentos4-paddlepaddle
GPU
PaddlePaddle: 3.2.0
CUDA: 12.9.1
NCCL: 2.27.6
cuDNN: 9.11.0.98
cuDNN 8: 8.9.7.29
Python: 3.11.6
OS: TencentOS 4
docker.io/tencentos/tencentos4-paddlepaddle
A PaddlePaddle 3.2.0 GPU Docker image based on TencentOS4 and CUDA 12.9. The image includes an MNIST handwritten digit recognition training example for quick environment verification.
tencentos4-tensorflow
GPU
TensorFlow: 2.20.0
CUDA: 12.9.1
NCCL: 2.27.6
cuDNN: 9.11.0.98
cuDNN 8: 8.9.7.29
Python: 3.11.6
OS: TencentOS 4
docker.io/tencentos/tencentos4-tensorflow
A TensorFlow 2.20.0 GPU Docker image based on TencentOS4 and CUDA 12.9. The image includes an MNIST handwritten digit recognition training example for quick environment verification.
tencentos4-sglang
GPU
CUDA:12.8
cuDNN 9: 9.11.0.98
cuDNN 8: 8.9.7.29
PyTorch:2.8.0
Torchvision:0.23.0
SGLang:0.5.2
sgl_kernel:0.3.9.post2
Python:3.11.6
OS: TencentOS 4
docker.io/tencentos/tencentos4-sglang
A SGLang 0.5.2 GPU Docker image based on TencentOS4 and CUDA 12.8. The image includes the official SGLang example code repository for quick environment verification and learning.
tencentos4-vllm
GPU
CUDA:12.8
cuDNN 9: 9.11.0.98
cuDNN 8: 8.9.7.29
PyTorch:2.8.0
Torchvision:0.23.0
vLLM:0.10.2
Python:3.11.6
OS: TencentOS 4
docker.io/tencentos/tencentos4-vllm
 
A vLLM 0.10.2 GPU Docker image based on TencentOS4 and CUDA 12.8. The image includes the official vLLM example code repository for quick environment verification and learning.
tencentos4-tensorrt-llm
GPU
CUDA:12.8
cuDNN 9: 9.11.0.98
cuDNN 8: 8.9.7.29
TensorRT-LLM:1.0.0
Python:3.12.2
OS: TencentOS 4
docker.io/tencentos/tencentos4-tensorrt_llm
 
A TensorRT-LLM 1.0.0 GPU Docker image based on TencentOS4 and CUDA 12.8. The image includes the official TensorRT-LLM example code repository for quick environment verification and learning.
tencentos4-transformer
tencentos4-deepspeed
GPU
Transformers : 4.56.2
deepspeed :0.16.8
NCCL: 2.27.6
CUDA: 12.9.1
cuDNN: 9.11.0.98
cuDNN 8: 8.9.7.29
Python: 3.11.6
OS: TencentOS 4
https://hub.docker.com/r/tencentos/tencentos4-tf-ds
A Transformers 4.56.2 and deepspeed 0.16.8 GPU Docker image based on TencentOS4 and CUDA 12.9. The image includes a pretrained model fine-tuned based on bert-base-uncased.
tencentos4-youtuagent
GPU
CUDA:12.8
cuDNN 9: 9.11.0.98
cuDNN 8: 8.9.7.29
Youtu-Agent:0.3
Python:3.12.12
OS: TencentOS 4
docker.io/tencentos/tencentos4-youtu
 
A Youtu-Agent GPU Docker image based on TencentOS4 and CUDA 12.8. The image includes the official Youtu-Agent example code repository for quick environment verification and learning.
tencentos4-browseruse
GPU
browseruse:0.9.5
google-chrome
Python:3.12.12
OS: TencentOS 4
https://hub.docker.com/r/tencentos/tencentos4-browser-use
A ready-to-use browser-use agent image built on TencentOS4 for quick environment deployment and learning.

Ajuda e Suporte

Esta página foi útil?

Você também pode entrar em contato com a Equipe de vendas ou Enviar um tíquete em caso de ajuda.

comentários

tencent cloud

TencentOS Server

TencentOS AI Container Image

Types of TencentOS AI Container Images

Features and Strengths

BaseOS Container Image

Framework Image

Ajuda e Suporte

Image type	Image Definition	Typical Scenarios
BaseOS Container Image	Provides a stable and lightweight TencentOS base container environment, making it easy for users to build custom images.	Suitable for users to build customized AI images or serve as the underlying foundation for their own applications.
Framework Container Image	Comes with nearly twenty mainstream AI frameworks preinstalled on top of the base image, covering various scenarios such as training and inference.	Suitable for users to directly perform model training and inference based on it.
LLM Container Image	Further integrates a complete toolchain and large model applications on top of the framework image, supporting out-of-the-box model services.	Suitable for users to directly deploy services and shorten the service launch cycle.

Image Name	Image address	Image description
TencentOS 4 minimal container image	docker.io/tencentos/tencentos4-minimal	The TencentOS 4 minimal container image is a minimal base image with full software package management features, using dnf as the package manager and coreutils as the Linux toolset.

Image Name	CPU/GPU	Version information	Image address	Image description
tencentos4-pytorch	GPU	PyTorch: 2.8.0 CUDA: 12.9.1 NCCL: 2.27.6 cuDNN 9: 9.11.0.98 cuDNN 8: 8.9.7.29 Python: 3.11.6 OS: TencentOS 4	docker.io/tencentos/tencentos4-pytorch	A PyTorch 2.8.0 GPU Docker image based on TencentOS4 and CUDA 12.9. The image includes the official PyTorch example code repository for quick environment verification and learning.
tencentos4-paddlepaddle	GPU	PaddlePaddle: 3.2.0 CUDA: 12.9.1 NCCL: 2.27.6 cuDNN: 9.11.0.98 cuDNN 8: 8.9.7.29 Python: 3.11.6 OS: TencentOS 4	docker.io/tencentos/tencentos4-paddlepaddle	A PaddlePaddle 3.2.0 GPU Docker image based on TencentOS4 and CUDA 12.9. The image includes an MNIST handwritten digit recognition training example for quick environment verification.
tencentos4-tensorflow	GPU	TensorFlow: 2.20.0 CUDA: 12.9.1 NCCL: 2.27.6 cuDNN: 9.11.0.98 cuDNN 8: 8.9.7.29 Python: 3.11.6 OS: TencentOS 4	docker.io/tencentos/tencentos4-tensorflow	A TensorFlow 2.20.0 GPU Docker image based on TencentOS4 and CUDA 12.9. The image includes an MNIST handwritten digit recognition training example for quick environment verification.
tencentos4-sglang	GPU	CUDA:12.8 cuDNN 9: 9.11.0.98 cuDNN 8: 8.9.7.29 PyTorch:2.8.0 Torchvision:0.23.0 SGLang:0.5.2 sgl_kernel:0.3.9.post2 Python:3.11.6 OS: TencentOS 4	docker.io/tencentos/tencentos4-sglang	A SGLang 0.5.2 GPU Docker image based on TencentOS4 and CUDA 12.8. The image includes the official SGLang example code repository for quick environment verification and learning.
tencentos4-vllm	GPU	CUDA:12.8 cuDNN 9: 9.11.0.98 cuDNN 8: 8.9.7.29 PyTorch:2.8.0 Torchvision:0.23.0 vLLM:0.10.2 Python:3.11.6 OS: TencentOS 4	docker.io/tencentos/tencentos4-vllm	A vLLM 0.10.2 GPU Docker image based on TencentOS4 and CUDA 12.8. The image includes the official vLLM example code repository for quick environment verification and learning.
tencentos4-tensorrt-llm	GPU	CUDA:12.8 cuDNN 9: 9.11.0.98 cuDNN 8: 8.9.7.29 TensorRT-LLM:1.0.0 Python:3.12.2 OS: TencentOS 4	docker.io/tencentos/tencentos4-tensorrt_llm	A TensorRT-LLM 1.0.0 GPU Docker image based on TencentOS4 and CUDA 12.8. The image includes the official TensorRT-LLM example code repository for quick environment verification and learning.
tencentos4-transformer tencentos4-deepspeed	GPU	Transformers : 4.56.2 deepspeed :0.16.8 NCCL: 2.27.6 CUDA: 12.9.1 cuDNN: 9.11.0.98 cuDNN 8: 8.9.7.29 Python: 3.11.6 OS: TencentOS 4	https://hub.docker.com/r/tencentos/tencentos4-tf-ds	A Transformers 4.56.2 and deepspeed 0.16.8 GPU Docker image based on TencentOS4 and CUDA 12.9. The image includes a pretrained model fine-tuned based on bert-base-uncased.
tencentos4-youtuagent	GPU	CUDA:12.8 cuDNN 9: 9.11.0.98 cuDNN 8: 8.9.7.29 Youtu-Agent:0.3 Python:3.12.12 OS: TencentOS 4	docker.io/tencentos/tencentos4-youtu	A Youtu-Agent GPU Docker image based on TencentOS4 and CUDA 12.8. The image includes the official Youtu-Agent example code repository for quick environment verification and learning.
tencentos4-browseruse	GPU	browseruse:0.9.5 google-chrome Python:3.12.12 OS: TencentOS 4	https://hub.docker.com/r/tencentos/tencentos4-browser-use	A ready-to-use browser-use agent image built on TencentOS4 for quick environment deployment and learning.