NVIDIA T4

Data center GPU based on NVIDIA's Turing architecture, announced in 2018 and designed for low-power inference workloads.

1 articleLast mentioned

The NVIDIA T4 is a data center GPU that NVIDIA announced in 2018. Built on the Turing architecture, it has 16GB of GDDR6 memory and Tensor Cores in a single-slot card with a 70W power draw, which makes it easy to install several per server. Launched under the name Tesla T4, it targets inference, video processing, and virtual desktop workloads rather than large-scale training.

It is widely available on major clouds such as Google Cloud and AWS and is known as the GPU allocated in the free tier of Google Colab. Although slower than newer GPUs, it is sufficient for loading open-weight models of a few billion parameters for inference and for small experiments, so it is frequently used in education.

This entry is based on AIPOST articles and widely known facts. If something is wrong, please send us a correction request.

Articles covering this entry

  • Pre-Training, Fine-Tuning and RL: Where Frontier LLMs Get Their QualityAI Performance ·

    …5B, a 500 million parameter open-weight model, can be loaded in a free Google Colab notebook on an NVIDIA T4 GPU through the Hugging Face Transformers pipeline.

    Read article

Found a mistake or missing an entry?

Wiki entries are based on AIPOST articles and widely known facts. If an entry is wrong, send us a correction request and we will check and fix it. Names from new articles are added to the wiki after the article is published.