Torch-TensorRT 1.0.0 #691

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Merged

narendasan merged 113 commits into master from torch_tensorrt_1.0

Nov 9, 2021

Collaborator

narendasan commented Nov 9, 2021

Description

This PR is to merge Torch-TensorRT (formally TRTorch) v1.0.0. Obviously this is a breaking change as all API namespaces have changed as a result of this PR. It also includes a number of features and changes

Turns on partial compilation by default
Changes the default expect input type given a model to be more in line with PyTorch
Workspace size defaults to 1GB for all platforms with GPUs newer than Maxwell (Maxwell defaults to 256MB)
The dockerfile has be redesigned to be much more minimal and robust

Type of change

Please delete options that are not relevant and/or add your own.

Bug fix (non-breaking change which fixes an issue)
New feature (non-breaking change which adds functionality)
Breaking change (fix or feature that would cause existing functionality to not work as expected)
This change requires a documentation update

Checklist:

My code follows the style guidelines of this project (You can use the linters)
I have performed a self-review of my own code
I have commented my code, particularly in hard-to-understand areas and hacks
I have made corresponding changes to the documentation
I have added tests to verify my fix or my feature
New and existing unit tests pass locally with my changes

inocsin and others added 30 commits

October 19, 2021 10:23


          feat: support setting input types of subgraph in fallback, handle Ten…

4778b2b

…sor type in evaluated_value_map branch in MarkOutputs

Signed-off-by: vcheungyi@163.com <vcheungyi@163.com>


          feat: support aten::conv1d and aten::conv_transpose1d

c8dc6e9

Signed-off-by: Ruoqian Guo <ruoqiang@nvidia.com>


          feat: support truncate_long_and_double in fallback subgraph input type

0bc3c05

Signed-off-by: inocsin <vcheungyi@163.com>


          merge from master

606d4de

Signed-off-by: inocsin <vcheungyi@163.com>


          feat: handle scalar type of size [] in shape_analysis

fca53ce

Signed-off-by: inocsin <vcheungyi@163.com>


          Merge remote-tracking branch 'upstream/master' into fix_conv1d

01417cd


          lower aten::conv1d and aten::conv_transpose1d to aten:convolution

b6d36aa

Signed-off-by: Ruoqian Guo <ruoqiang@nvidia.com>


          refactor!: API level rename

483ef59

BREAKING CHANGE: This commit renames the namespaces of all
TRTorch/Torch-TensorRT APIs. Now torchscript specific functions
are segregated into their own torch_tensorrt::torchscript /
torch_tensorrt.ts namespaces. Generic utils will remain in the
torch_tensorrt namespace. Guidance on how to port forward will follow in
the next commits


          chore: add warning message when truncating graph input type of kLong/…

3037f75

…kDouble

Signed-off-by: inocsin <vcheungyi@163.com>


          refactor(//py): Redo Python tests for new API

a9ab1b4

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          refactor: Apply linting

b98eac4

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          feat: Update documentation with new library name Torch-TensorRT

e5f96d9

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          refactor(//docker): Update docker container for new API

05b8376

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          refactor! : Update README.md with the new library name Torch-TensorRT

8d9722d

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          Merge branch 'torch_tensorrt_rebrand' into docs_refactor

ded366d


          chore: More README updates

e7bbc2d

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          chore: Doc updates

0a2405f

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          fix: Fix plugin registration macro

8afab22

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          refactor: Finishing porting namespace

0c42277

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          chore: Rebase with torch_tensorrt_rebrand branch

fb59b80

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          Merge branch 'docs_refactor' into 'torch_tensorrt_rebrand'

05387f9

feat: Update documentation with new library name Torch-TensorRT

See merge request adlsa/TRTorch!1


          refactor: Rebranding complete

715a0ce

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          refactor: Realign py api

493c19f

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          refactor: Update example names

d872d02

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          refactor: update docs

65ffaef

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          fix: Update notebooks with new library name Torch-TensorRT

8274fd9

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          chore: Update Docker instructions in main README.md

af13839

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          chore: Update README C++ code snippet

c4723a9

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          Merge branch 'notebook' into 'main'

fix: Update notebooks with new library name Torch-TensorRT

See merge request nvidia/trtorch/Torch-TensorRT-Preview!1


          Merge branch 'readme' into 'main'

4d606bc

chore: Update Docker instructions in main README.md

See merge request nvidia/trtorch/Torch-TensorRT-Preview!2

narendasan and others added 12 commits

November 8, 2021 15:17


          Merge branch 'ncomly-torch_tensorrt_rebrand-patch-42370' into 'releas…

93e4c4a

…e/1.0'

Update README.md with new API & pointing to NGC container

See merge request adlsa/TRTorch!21


          doc: Update installation instructions based on ABI and platforms

d06eba7

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          chore: Update doc

c3d648b

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          chore: Minor update

8c4224b

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          Merge branch 'readme' into 'release/1.0'

38685a0

doc: Update installation instructions based on ABI and platforms

See merge request adlsa/TRTorch!22


          chore: Fix python jetson installation instruction

62bdc45

Signed-off-by: Dheeraj Peri <peri.dheeraj@gmail.com>


          Merge branch 'readme' into 'release/1.0'

chore: Fix python jetson installation instruction

See merge request adlsa/TRTorch!23


          chore: Updates to documentation

afa97b9

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          chore: Regen docs for final review

fc6d5ff

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          chore: update list of supported ops

8b1221b

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          refactor: Apply linting

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>


          chore: Docs for v1.0.0

7ab76bc

Signed-off-by: Naren Dasan <naren@narendasan.com>
Signed-off-by: Naren Dasan <narens@nvidia.com>

narendasan requested a review from peri044

November 9, 2021 06:07

github-actions bot added component: api [Python] component: api [C++] component: build system component: conversion component: converters component: core component: evaluators component: lowering component: partitioning component: runtime component: tests documentation labels

narendasan requested review from ncomly-nvidia and andi4191

November 9, 2021 06:12

narendasan force-pushed the torch_tensorrt_1.0 branch from 660d9b6 to 7ab76bc Compare

November 9, 2021 06:14

narendasan merged commit 3a98a8b into master

andi4191 deleted the torch_tensorrt_1.0 branch

August 10, 2022 22:23

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

component: api [C++] component: api [Python] component: build system component: conversion component: converters component: core component: evaluators component: lowering component: partitioning component: runtime component: tests documentation