Wikitopia
FasterTransformer
productinference_framework
Try in Playground →
Overview
Developed byNVIDIA
LicenseApache License 2.0
Open source✓ Open Source
Use caseAccelerating Transformer model inference
Knowledge graph stats
Claims31
Avg confidence92%
Avg freshness99%
Last updatedUpdated 5 days ago
Trust distribution
100% unverified
Governance
Not assessed
Contribute governance data →

FasterTransformer

product

NVIDIA's CUDA-based library for accelerated transformer model inference with various optimization techniques.

Compare with...

primary use case

ValueTrustConfidenceFreshnessSources
Accelerating Transformer model inference○UnverifiedHighFresh1
Transformer model inference optimization○UnverifiedHighFresh1
High-performance transformer model inference acceleration○UnverifiedHighFresh1
GPU-accelerated transformer model inference○UnverifiedHighFresh1
High-performance transformer model inference optimization○UnverifiedHighFresh1
accelerated transformer model inference○UnverifiedHighFresh1

pricing model

ValueTrustConfidenceFreshnessSources
free○UnverifiedHighFresh1

open source

ValueTrustConfidenceFreshnessSources
true○UnverifiedHighFresh1

developed by

ValueTrustConfidenceFreshnessSources
NVIDIA○UnverifiedHighFresh1

maintained by

ValueTrustConfidenceFreshnessSources
NVIDIA○UnverifiedHighFresh1

integrates with

ValueTrustConfidenceFreshnessSources
CUDA○UnverifiedHighFresh1
Triton Inference Server○UnverifiedHighFresh1
PyTorch○UnverifiedHighFresh1
TensorRT○UnverifiedModerateFresh1
TensorFlow○UnverifiedModerateFresh1

written in

ValueTrustConfidenceFreshnessSources
C++○UnverifiedHighFresh1
CUDA○UnverifiedHighFresh1

supports model

ValueTrustConfidenceFreshnessSources
BERT○UnverifiedHighFresh1
GPT○UnverifiedHighFresh1
T5○UnverifiedHighFresh1

based on

ValueTrustConfidenceFreshnessSources
C++/CUDA○UnverifiedHighFresh1
C++○UnverifiedHighFresh1
CUDA kernels○UnverifiedModerateFresh1

requires

ValueTrustConfidenceFreshnessSources
NVIDIA GPU○UnverifiedHighFresh1
CUDA○UnverifiedHighFresh1

license type

ValueTrustConfidenceFreshnessSources
Apache License 2.0○UnverifiedHighFresh1

competes with

ValueTrustConfidenceFreshnessSources
TensorRT○UnverifiedModerateFresh1
Hugging Face Transformers○UnverifiedModerateFresh1

alternative to

ValueTrustConfidenceFreshnessSources
Hugging Face Transformers○UnverifiedModerateFresh1

Alternatives & Similar Tools

TensorRT
competes with
Compare →
Hugging Face Transformers
competes with
Compare →
Hugging Face Transformers
alternative to
Compare →

Commonly Used With

CUDATriton Inference ServerPyTorchTensorRTTensorFlow

Related entities

Claim count: 31Last updated: 4/5/2026Edit history