ProductTesla T4 / T4 GPU
NVIDIA T4
A data-center GPU from NVIDIA designed for inference workloads. Based on the Turing architecture, it features 16 GB GDDR6 memory and 320 Tensor Cores. Characterized by low power consumption and broad cloud availability, it is widely used for AI inference acceleration.
Timeline (last 90 days)
Sep 16
The NVIDIA T4 is based on the Turing architecture, with 16GB GDDR6 VRAM and 320 Tensor Cores per card.
Unverified50%
Sep 16
A 70B-parameter model requires approximately 35–40GB of VRAM after Q4 quantization, which a single T4 cannot handle, but dual T4s can basically satisfy.
Unverified50%