Seamless analysis of your PyTorch models (RAM usage, FLOPs, MACs, receptive field, etc.)
APACHE-2.0 License
Published by frgfm about 2 years ago
This release fixes the support of some layers and improves the project quality.
Note: torchscan 0.1.2 requires PyTorch 1.5 or higher.
It was time to update the documentation, thus the theme was changed from Read the Docs to Furo (#61)
This comes with nice features like dark mode and edit button!
Getting a clean, maintainable code base is one of the key catalyst aspects for an OSS project. As such, a few improvements have been made:
requirements.txt
to pyproject.toml
(#58, #59)The master
branch was renamed into main
(#60), so if you were installing from source a specific branch name, make sure to update this reference!
Full Changelog: https://github.com/frgfm/torch-scan/compare/v0.1.1...v0.1.2
Published by frgfm about 4 years ago
This release adds support of more modules by the crawler and enables receptive field computation for highway nets.
Note: torchscan 0.1.1 requires PyTorch 1.1 or newer.
In-hook information extraction for supported torch.nn.Module
New
torch.nn.Module
: Identity
, Linear
, Identity
, ReLU
, ELU
, LeakyReLU
, ReLU6
, Tanh
, Sigmoid
, _ConvTransposeNd
, _ConvNd
, _BatchNorm
, _MaxPoolNd
, _AvgPoolNd
, _AdaptiveMaxPoolNd
, _AdaptiveAvgPoolNd
, Dropout
(#21).Fixes
## Crawler
Module hooking agent
New
receptive_field
in the summary
function (#21)Verifications of the package well-being before release
New
torchscan.modules
(#21)Online resources for potential users
Improvements
Fixes
Other tools and implementations
Published by frgfm over 4 years ago
This release adds a module crawler to pick up relevant inference information.
Note: torchscan 0.1.0 requires PyTorch 1.1 or newer.
In-hook information extraction for supported torch.nn.Module
New
torch.nn.Module
: Identity
, Linear
, Identity
, ReLU
, ELU
, LeakyReLU
, ReLU6
, Tanh
, Sigmoid
, _ConvTransposeMixin
, _ConvNd
, _BatchNorm
, _MaxPoolNd
, _AvgPoolNd
, _AdaptiveMaxPoolNd
, _AdaptiveAvgPoolNd
, Dropout
(#1, #6, #7).Python process information related
New
get_process_gpu_ram
to retrieve GPU RAM usage of the current Python process (#1).## Crawler
Module hooking agent
New
crawl_module
to store all module information in a python dict
(#1, #6)summary
for high-level console-printed information (#1)Verifications of the package well-being before release
New
torchscan.modules
(#1 , #6, #7)torschscan.process
(#1, #6)torschscan.crawler
(#6)torschscan.utils
(#1, #6)Online resources for potential users
New
Other tools and implementations
to generate a string output from the
crawl_module` returned dictionary (#1).aggregate_info
to aggregate crawl_module
output to a specific depth (#1).scripts/benchmark.py
to display crawl_module
information on all torchvision
classification models (#1 )Notes: upon the next torch
release, _ConvTransposeMixin
will be renamed to _ConvTransposeNd