Technology Encyclopedia Home >How to quickly get started with Tencent Cloud TI platform?

How to quickly get started with Tencent Cloud TI platform?

Created on 2025-08-01 19:39:08
97

To quickly get started with the Tencent Cloud TI Platform (Tencent Intelligent Platform), follow these steps:

1. Understand the TI Platform

The Tencent Cloud TI Platform is a suite of AI and big data tools designed to help developers and enterprises build, train, and deploy machine learning models efficiently. It includes services like TI-ONE (AI training platform), TI-EMS (model management), and TI-Accel (accelerated computing).

2. Access the Platform

  • Log in to the Tencent Cloud Console.
  • Search for "TI Platform" or navigate to "Artificial Intelligence" > "TI Platform".

3. Key Services & How to Use Them

TI-ONE (AI Training Platform)

  • Use Case: Train custom machine learning models (e.g., image recognition, NLP).
  • How to Start:
    • Create a new workspace.
    • Upload datasets (supporting formats like CSV, TFRecord).
    • Choose a pre-built algorithm (e.g., TensorFlow, PyTorch) or bring your own.
    • Configure compute resources (GPU/CPU clusters).
    • Start training and monitor progress.

TI-EMS (Model Management & Deployment)

  • Use Case: Deploy trained models as APIs for real-time inference.
  • How to Start:
    • Import models from TI-ONE or upload directly.
    • Configure auto-scaling and endpoint deployment.
    • Test the API using the built-in dashboard.

TI-Accel (Accelerated Computing)

  • Use Case: High-performance computing for large-scale AI workloads.
  • How to Start:
    • Select GPU-accelerated instances (e.g., NVIDIA T4, V100).
    • Optimize training with distributed computing support.

4. Example Workflow (Image Classification)

  1. Upload Dataset: A labeled image dataset (e.g., cats vs. dogs).
  2. Train Model: Use TI-ONE with a PyTorch-based CNN model.
  3. Deploy Model: Export to TI-EMS and expose as a REST API.
  4. Inference: Call the API to classify new images in real time.

5. Recommended Tencent Cloud Services for AI/Big Data

  • Cloud Storage (COS): Store datasets efficiently.
  • TencentDB: Manage structured data for training.
  • Tencent Cloud Edge Computing: Deploy AI at the edge for low-latency inference.

Start with TI-ONE’s pre-built templates (e.g., sentiment analysis, object detection) to accelerate development. For large-scale needs, leverage TI-Accel’s GPU clusters.

For documentation, visit Tencent Cloud TI Platform Docs.