gptkbp:instanceOf
|
hardware accelerator
|
gptkbp:announced
|
gptkb:Google_I/O_2016
|
gptkbp:architecture
|
ASIC
|
gptkbp:availableOn
|
gptkb:Cloud_TPU
gptkb:Google_Cloud_Platform
gptkb:Edge_TPU
|
gptkbp:competitor
|
gptkb:NVIDIA_GPUs
gptkb:AMD_GPUs
FPGAs
|
gptkbp:developedBy
|
gptkb:Google
|
gptkbp:energyEfficiency
|
true
|
gptkbp:firstReleased
|
2016
|
gptkbp:fullName
|
gptkb:Tensor_Processing_Units
|
gptkbp:generation
|
v3
v1
v2
v4
|
gptkbp:hasComponent
|
gptkb:High_Bandwidth_Memory
gptkb:Interconnect
gptkb:Matrix_Multiply_Unit
|
https://www.w3.org/2000/01/rdf-schema#label
|
TPUs
|
gptkbp:location
|
data centers
|
gptkbp:openSource
|
gptkb:TensorFlow_Lite_for_Edge_TPU
|
gptkbp:optimizedFor
|
matrix multiplication
neural network training
neural network inference
|
gptkbp:platform
|
gptkb:TensorFlow
|
gptkbp:relatedTo
|
gptkb:Google_Cloud
gptkb:TensorFlow
gptkb:Google_AI
|
gptkbp:supports
|
computer vision
natural language processing
distributed training
large language models
|
gptkbp:supportsPrecision
|
bfloat16
float32
int8
|
gptkbp:usedBy
|
gptkb:Google_Assistant
gptkb:Google_Photos
gptkb:Google_Search
gptkb:Google_Translate
|
gptkbp:usedFor
|
gptkb:artificial_intelligence
gptkb:machine_learning
deep learning
|
gptkbp:usedIn
|
gptkb:TPU_Pods
|
gptkbp:website
|
https://cloud.google.com/tpu
|
gptkbp:bfsParent
|
gptkb:AI_Infrastructure
|
gptkbp:bfsLayer
|
5
|