Repository Issue Activity (beta)

#1419 Support for AMD RDNA 4 (GFX1200)

#1334 vLLM doesnt work on DGX spark

#1335 nvidia-ml-py not supported on DGX Spark

#1281 Update raw SQL queries to use SQLAlchemy

#2295 Public IPs not freed when using Azure provider

#1282 Move all user DB code somewhere in the DB directory

#2195 UI for managing experiments needs to be cleaned up

#760 Support using checkpoints as inference adaptors

#1120 Feature: Support fine-tuning for Z-Image architecture

#918 Update docs to explain how to develop frontend in cloud mode

#1446 Slurm tasks stay in LAUNCHING state after success or failure

#2056 CLI: support --credentials-file for provider add to keep secrets out of argv

#2040 Reduce payload size for jobs list

#2029 task queue succeeds against a non-existent experiment_id (server creates orphan job)

#2002 Update autoresearch skill to take more notes as it goes

#1997 lab provider list: table wrapping makes provider names ambiguous; agents pick the wrong name silently

#2004 Add total jobs count in the jobs score chart

#1981 Remote SkyPilot jobs falsely marked FAILED during cluster provisioning

#1999 Need UI in webapp to view "description" per job. Maybe a tiny notebook icon that you can hover over for a tooltip and click on for a full modal that shows the full descirption for that job

#1988 lab.finish: score field write silently fails intermittently while completion_details persists

#1969 Destructive job endpoints (`/jobs/delete/...`) use HTTP GET, should be DELETE

#1986 lab job discard: TypeError: put() got an unexpected keyword argument 'json'

#1971 Document `lab.finish(score=...)` — it's a dict, not a scalar

#1903 No CLI command to restart or stop the TLab server

#1968 Job delete endpoints return 200 OK without actually deleting (silently swallowed exception)