Skip to content

User Guides

An Introduction To Tripy
Using the Compiler
Quantization
Custom Operations

API Reference

Compiling Code
config
dtype
device
Exception
- TripyException
logger
Module
- Sequential
Operations
- Functions
  - all
  - allclose
  - any
  - argmax
  - argmin
  - avgpool
  - cast
  - concatenate
  - copy
  - cos
  - cumsum
  - equal
  - exp
  - expand
  - flatten
  - flip
  - gather
  - gelu
  - log
  - masked_fill
  - max
  - maximum
  - maxpool
  - mean
  - min
  - minimum
  - outer
  - pad
  - permute
  - plugin
  - prod
  - relu
  - repeat
  - reshape
  - resize
  - rsqrt
  - sigmoid
  - silu
  - sin
  - softmax
  - split
  - sqrt
  - squeeze
  - stack
  - sum
  - tanh
  - topk
  - transpose
  - unsqueeze
  - var
  - where
- Initializers
  - arange
  - full
  - full_like
  - iota
  - iota_like
  - ones
  - ones_like
  - tril
  - triu
  - zeros
  - zeros_like
- Modules
  - BatchNorm
  - Conv
  - ConvTranspose
  - Embedding
  - GroupNorm
  - InstanceNorm
  - LayerNorm
  - Linear
- Quantization
  - dequantize
  - quantize
Tensor
- DimensionSize
types

Developer Guides

Architecture
Adding New Operations
Debugging MLIR-TensorRT

Compiling Code¶

compile
DimensionInputInfo
Executable
InputInfo
Stream

Custom Operations

Copyright © 2024, NVIDIA

Made with Sphinx and @pradyunsg's Furo