* Add `g6e_xlarge_task`, `g6e_2xlarge_task`, `g6e_4xlarge_task`, `g6e_8xlarge_task`, `g6e_12xlarge_task`, `g6e_16xlarge_task`, and `g6e_24xlarge_task` to request G6e GPU instances.
Updated
* Update latch base image to 'cb01-main'
2.55.0
Changed
* `latch dockerfile` now accepts explicit options for files containing * apt deps * conda deps * r deps * pip deps * pyproject.toml / setup.py * .env * `latch dockerfile` no longer does implicit dependency inferrence based on the presence of special files in the project root directory - now all dependency files need to be passed explicitly in the command line.
2.54.10
Changed
* Bump ephemeral storage limits for V100 GPU pods
2.54.9
Fixed
* Reduced cpu/memory requests to ensure large GPU pods are schedulable