Statements (13)
| Predicate | Object |
|---|---|
| gptkbp:instance_of |
gptkb:document
|
| gptkbp:developed_by |
gptkb:NVIDIA
|
| https://www.w3.org/2000/01/rdf-schema#label |
TF32
|
| gptkbp:improves |
AI training performance
|
| gptkbp:introduced_in |
gptkb:Ampere_architecture
|
| gptkbp:is_compatible_with |
gptkb:CUDA
|
| gptkbp:is_optimized_for |
deep learning tasks
|
| gptkbp:is_related_to |
tensor cores
|
| gptkbp:provides |
higher throughput than FP32
|
| gptkbp:supports |
mixed precision training
|
| gptkbp:used_in |
gptkb:NVIDIA_GPUs
|
| gptkbp:bfsParent |
gptkb:NVIDIA_A100_Tensor_Core_GPU
|
| gptkbp:bfsLayer |
5
|