We've released a number of improvements and bugfixes for Gradient!
Gradient Notebooks now autosave by default improvement
We've improved the autosave functionality of notebooks! Whereas before only
.ipynb files would save automatically, we now provide autosave functionality for all filetypes within notebooks.
Notebooks running on free GPU instances can now be private on Pro or Growth subscriptions improvement
If you're on the Gradient Pro or Growth plan, notebooks that run on Free GPU instances can now be made private.
PyTorch container updated to version 1.10, TensorFlow container updated to 2.6.0
We've updated both PyTorch and TensorFlow default containers in Gradient Notebooks to their latest stable release versions. The new runtimes are now available in the Gradient console.
- Gradient Deployments are now able to pull from models registered in Gradient
- Overall GPU capacity has increased after addressing an issue related to read-only filesystems used by Gradient Notebooks
- We fixed a bug in the notebook create menu that sometimes caused the
Workspace URLfield not to update when selecting a new runtime
- We fixed a bug in notebooks that sometimes caused deleted files to linger in the file management pane
- We fixed a bug in notebooks that caused an empty file to be added to new directories
- We fixed a bug that sometimes generated duplicate and triplicate notifications when switching teams