Original Source Here
For more customization examples head over to the documentation page.
Where is AutoML or Hyperparamer optimization?
There are two ways to build and tune your model hyperparameter in Gradsflow.
Gradsflow provides an AutoModel class (experimental 🚨) which is very similar to the Model class that you just saw. You can register the hyperparameters which you want to tune with the Tuner class and with minimal code changes you will be ready to train your model with Hyperparameter optimization mode.
In this example, Tuner will optimize for cnn1 and cnn2 architecture and adam and SGD optimizer.
This is how you can do hyperparameter tuning using the Model and AutoModel class.
The other way to automatically build and train models are AutoTask. Autotask provides low code model building and training classes for various tasks like Image classification, Text classification & Text Summarization. It is powered by PyTorch Lightning Flash.
Find my talk at DeepSource.io Hacktoberfest 2021 Kickoff 🎉
📚 To read more about Gradsflow you can visit the documentation page https://docs.gradsflow.com/en/latest. You can find examples for training Auto Image classification, Auto Text Summarization, Pix2Pix GAN training and HuggingFace model training.
🙏🏻 Gradsflow takes inspiration for APIs & leverages some of the great OSS libraries including Ray️, HuggingFace Accelerate, PyTorch Lightning TorchMetrics, Keras & fast.ai.
🌟 Like this Project? Give us a star on GitHub
Have questions? Join our Slack community 🧡 to chat with us.
Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot