archinetai/surgeon-pytorch
A library to inspect and extract intermediate layers of PyTorch models.
This tool helps machine learning engineers and researchers understand and modify complex PyTorch models without altering the original code. You can easily see the data (tensors) flowing through specific layers or even cut out a section of a model to run independently. This allows for deep analysis, visualization, and debugging of neural network components.
473 stars. No commits in the last 6 months. Available on PyPI.
Use this if you need to understand the internal workings of a PyTorch model, visualize intermediate outputs, or extract specific parts of a model for separate use or debugging without changing its source code.
Not ideal if your PyTorch model uses dynamic execution graphs (e.g., control flow like `for` loops or `if` statements that depend on input data) and you intend to use the `Extract` feature.
Stars
473
Forks
16
Language
Python
License
MIT
Category
Last pushed
May 12, 2022
Commits (30d)
0
Dependencies
2
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/archinetai/surgeon-pytorch"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
MaximeVandegar/Papers-in-100-Lines-of-Code
Implementation of papers in 100 lines of code.
kk7nc/RMDL
RMDL: Random Multimodel Deep Learning for Classification
OML-Team/open-metric-learning
Metric learning and retrieval pipelines, models and zoo.
miguelvr/dropblock
Implementation of DropBlock: A regularization method for convolutional networks in PyTorch.
PaddlePaddle/models
Officially maintained, supported by PaddlePaddle, including CV, NLP, Speech, Rec, TS, big models...