studioml / studio

Studio: Simplify and expedite model building process

https://studio.ml

studioml/studio Issues

Support user-defined mutable artifacts for experiment.
Updated 3 years ago
Support local task queues for external evaluation workers.
Updated 3 years ago
Detect bad RabbitMQ credentials on queue creation.
Closed 3 years ago
Put package structure into StudioML code base.
Closed 3 years ago
Improve KeyboardInterrupt handling.
Closed 3 years ago
Remove use of GNU-licensed Python libraries.
Closed 3 years ago
Improve performance and management of workload storage handlers.
Closed 3 years ago1
Support for persistent workload queues.
Updated 3 years ago
Switch to using Python libraries for tar/untar file processing.
Closed 3 years ago
Fix HTTP artifacts database provider.
Closed 3 years ago
Allow studio/completion service to clean up after itself better
Updated 4 years ago
Change boto3 based functions for Google Cloud Storage compatability
Updated 4 years ago
Allow external dependencies policies in StudioML.
Closed 4 years ago1
Factor out CompletionService into its own library
Closed 4 years ago1
Remove studioml dependency on Keras and TensorFlow.
Closed 4 years ago1
Allow studioml to use external experiment payload builders.
Closed 4 years ago
Improve error reporting for artifacts upload/download.
Closed 4 years ago
Clean up studioml dependencies.
Closed 4 years ago
Fix experiment setup for DB provider and Artifact store.
Closed 4 years ago
Refactor studioml code to prepare splitting out Completion Service to different implementation.
Closed 4 years ago
Specify experiment workload dependencies explicitly.
Updated 4 years ago
Cleanup studioml Python dependencies.
Closed 4 years ago
Lock boto3 dependency version.
Closed 4 years ago
Better log volume management.
Closed 4 years ago
Provide experiment workload digital signing.
Updated 4 years ago
Potential race condition between Studio API and RabbitMQ event loop.
Closed 4 years ago
Add bibtex file
Closed 4 years ago
RMQ queues left undeleted after experiment run.
Closed 4 years ago
Local queue should not timeout on getting new message.
Closed 4 years ago1
Provide experiment workload encryption.
Updated 4 years ago
Reschedule failed workloads in local StudioML configuration.
Updated 4 years ago
Log reprinter should stop when Completion service is shutdown.
Closed 4 years ago
Local tasks queue needs "deleted" state.
Closed 4 years ago
All message queue (sub)classes must implement shutdown() API.
Closed 4 years ago
Completion service should shutdown its transport jobs queue.
Closed 4 years ago
studio tends to pull in tensorflow==1.14.0, should pull in at least 1.15.2
Closed 4 years ago1
studioml wants PyYAML==5.2 as dependency, but gives a warning message
Closed 4 years ago1
studio tends to uninstall pip==20.0.2 in favor of pip==19.0.3
Closed 4 years ago1
Consider updating to latest version of pika
Closed 4 years ago1
Completion Service file spec interface should be more robust
Updated 4 years ago1
Consistency between local and distributed completion service file specifications
Updated 4 years ago
Studio Ml future roadmap
Closed 4 years ago2
Implement local storage/database mode for studioml users.
Closed 4 years ago
pip python module needs version control
Updated 4 years ago
Alternate endpoint handling in local Python Completion Service execution
Closed 4 years ago
Disable the use of Python 2, Mitigate the tensorflow CVE for versions prior to 1.15.2
Closed 4 years ago
Python runtime versions not explicit enough
Closed 5 years ago
Update pika dependency to 0.13.1
Closed 5 years ago
_metadata artifacts are not updated
Closed 5 years ago
Per artifact credentials
Updated 5 years ago