Release 1.1.0

Release Notes

New Feature: profile_model for Convenient Model Profiling

• Default Input Shape: (1, 3, 224, 224)

How to Get Model Information with timm:

from docsaidkit.torch import profile_model
from pprint import pprint

pprint(profile_model('resnet50'))

How to Profile Your Custom Model:

import torch.nn as nn
from docsaidkit.torch import profile_model
from pprint import pprint

model = nn.Conv2d(3, 64, 3, 1, 1)

pprint(profile_model(model))

This update provides an easy way to obtain detailed information about models, including those from the timm library and your own custom models.

Full Changelog: 1.0.1...1.1.0