0.17.6
Changelog
- a7806b5 chore: bump version: 0.17.6-rc6 -> 0.17.6
- 48451b6 chore: fix import ordering
- d9a1257 docs: add release notes for 0.17.6 (#3475)
- 47a17f6 chore: bump version: 0.17.6-rc5 -> 0.17.6-rc6
- 7a8132a fix: make systemd socket activation actually work (#3459)
- b6d82d1 chore: bump version: 0.17.6-rc4 -> 0.17.6-rc5
- 6e65020 fix: Allow allocations array to be empty without null value (#3465)
- 79ca9dd fix: update timeago casing for taskcards on dashboard (#3434)
- f40314e chore: bump version: 0.17.6-rc3 -> 0.17.6-rc4
- 247c288 fix: avoid terminating profiler streaming [DET-6459] (#3453)
- 90a385e chore: bump version: 0.17.6-rc2 -> 0.17.6-rc3
- bb43d69 fix: revert to prior slot utilization logic for static agents (#3451)
- 75972b3 chore: bump version: 0.17.6-rc1 -> 0.17.6-rc2
- 34e2e31 chore: better telemetry (#3271)
- bdd5c23 ci: run releases for tags with new proper SemVer format
- c87dcf4 chore: bump version: 0.17.6-rc0 -> 0.17.6-rc1
- 49c03b4 chore: bump version: 0.17.6-dev0 -> 0.17.6-rc0
- e7ed8e9 chore: lock api state for backward compatibility check
- 1885905 fix: Add allocation state to db test object (#3427)
- 4dcf325 feat: allow podSpec env variables (#3431)
- 25616ac feat: adjust job priority and weight through job queue (#3411)
- c3c4df9 feat: Add /tasks/:task_id endpoint to GRPC API [DET-6354] [DET-6355] (#3360)
- db71ac4 docs: announce deprecation of pbt (#3407)
- 01da997 feat: pass metrics to simple reducer in original order (#3405)
- 3956804 fix: show correct total gpu capacity [DET-3733] (#3385)
- b0f5458 chore: bump env images for security. (#3415)
- 4f2ced6 fix: address experiment name going out of sync with db (#3414)
- dd733ef fix: avoid Can't pickle local object in TestPIDServer (#3393)
- 3dcdbc8 fix: add missing fields to allocation query and tests to prevent future bugs (#3398)
- c1d5db0 ci: fix flake in provisioner unit test (#3409)
- f082d65 chore: update unreleased manage job modal (#3374)
- cd64b92 chore: make mypy happy with requests wrapper (#3408)
- d0da869 fix: fix a conditional render loop (#3394)
- c3e2d3a chore: bumpenvs for updated base AMIs (#3404)
- c6f2318 chore: get gov images in refresh-ubuntu-amis.py (#3399)
- 7f504e7 ci: make checkpoint gc tests actually wait for gc (#3403)
- 142f599 fix: set gc-policy broken [DET-6373] (#3391)
- 47b2375 fix: fix forked experiments missing username in memory (#3392)
- 9a1e811 feat: add systemd socket activation support to the master (#3366)
- 875d673 DET-6361 - update docs (#3386)
- e9ad053 chore: force github.com/containerd/containerd upgrade (#3381)
- 7e822f5 chore: fix default format selection and enum loading in cli (#3384)
- dc511af chore: write our own swagger bindings (#3361)
- 9a5c1f5 fix: Fix sphinx-build parsing bug (#3376)
- 3bc0957 fix: stop re-rendering loops and throw the appropriate errors for continue trial modal [DET-6368] (#3378)
- 66378a4 chore: bump github.com/labstack/echo/v4 dependencies to address dependabot (#3354)
- 38f7775 fix: fix webui full config edit in notebook modal (#3373)
- 2debbcf chore: bump docker and k8s dependencies (#3352)
- b3a34ba docs: address onboarding gaps (#3122)
- a996243 ci: stop testing EOL python (#3377)
- e6ae62a chore: update github pr template (#3365)
- 0f45c4a fix: stop profiler spinner when terminal [DET-6326] (#3325)
- ceb537d fix: negative slots per agent [DET-6357] (#3342)
- 31549d8 fix: default shell/cmd slots should be 1. (#3369)
- 268035b chore: try to sidestep race in use of check_if_string_present_in_trial_logs test helper (#3367)
- eaeb658 chore: bump goreleaser (#3345)
- 6996598 chore: image updates: bump all, add ROCm image. (#3363)
- c81a7a9 ci: unflake master
IdleTimeoutWatcher
test. (#3364) - 4b76543 chore: fix small data race found by go build --race (#3359)
- ddda7a4 fix: purge model.ExperimentConfig (#3362)
- 1f4898c feat: experimental ROCm support. [DET-6285] (#3282)
- 611947c ci: only install yq with snap. (#3355)
- a0a5a8c chore: fix trial log readability (#3356)
- 34698e6 chore: AdvancedSearcher->Searcher (#3339)
- 98c69ad chore: More info for test failure [DET-6347] (#3353)
- c1a1e30 chore: add trial log dump for test assertion failure (#3336)
- 5241759 fix: handle calls to old command endpoints, make it harder to crash cmd managers [DET-6336] (#3315)
- 906ea4b fix: Convert all experiment and job states to labels (#3351)
- c9d43f9 chore: bump test-e2e go version (#3344)
- 95999f9 fix: fix incorrect preemption status report from Kubernetes RP (#3330)
- 1f8eaa9 chore: rename Generic API to Core API [DET-6243] (#3310)
- c266f8e ci: print trial logs in more failure cases (#3333)
- 61b309f fix: don't allow allocations to take actions with unreceived cancellations (#3326)
- 2e23499 fix: small bug in error log (#3332)
- 377560e feat: support agent on Apple Silicon without Rosetta (#3328)
- 9b31a1b feat: add config option for Tensorboard (#3319)
- 46d5627 refactor: stop experiment modal [DET-6325] (#3307)
- 12fce38 ci: update to python3.7. (#3316)
- 30397f2 chore: unpin google-cloud dependencies. (#3320)
- 4632cbe chore: simplify job queue state tracking (#3302)
- 7ff1f8e fix: collect system metrics from all agents (#3313) [DET-6332]
- 355dcb8 chore: workaround upstream torch bug (#3321)
- 0c12fe3 test: store and report webui test results (#3248)
- a2cede9 feat: add prometheus endpoint for internal Determined state mappings [DET-5890] (#3258)
- 7d9714e chore: update can-i-use browserlist (#3317)
- f4fc7bb chore: remove hvd_config usage [DET-6220] (#3210)
- cb139a1 fix: pull logs [DET-6335] (#3308)
- ca96c77 feat: add wall clock time, tests to get trial API [DET-6226] (#3311)
- 997dd8f fix: preview search (#3309)
- 43ccbea fix: improve CPU core count parsing on agents with CPU slots. (#3304)
- 9b80a57 chore: clean up code owners (#3312)
- dd3edaa docs: fix image formatting in 0.17.5 release notes (#3305)
- f9dd54a chore: bump version: 0.17.5-dev0 -> 0.17.6-dev0
- 09ca94f docs: add release notes for 0.17.5 (#3299)
- 09c049c chore: update job queue title and navigation entry (#3303)
- c189054 feat: Allow new AWS instances to be specified [DET-6327] (#3296)
- 8152a90 chore: reuse ordering logic between k8 and priority schedulers (#3301)
- 5716cf2 chore: reorganize how endpoints are queried in jobs page (#3298)
- 50be69e fix: update craco config to have webpack use the ify-loader for plotly imports (#3294)
- 54a37e3 fix: fix experiment active state check in webui (#3295)
- a2bdf72 fix: increase CircleCI resource class for React builds (#3297)
- e06e307 fix: open job queue task links in a new tab (#3293)
- 26b6d26 fix: pass get_trials sort parameters to REST (#3291)
- 60b2981 feat: set up, read, and visualize job queue (#3231)
- 6686b5f fix: update use in notebook code snippet [DET-6305] (#3288)
- 9707b82 Fix: send activate param from WebUI to API (#3290)
- 6432ca9 fix:
det model list-versions --json
(#3292) - ded1d76 chore: avoid progress rendering for tasks card in some cases (#3289)
- f4302d8 fix: add visual indicators that you can't edit an archived model [DET-6280] (#3279)
- f38996a refactor: customized timeago [DET-6244] (#3283)
Docker images
docker pull determinedai/determined-master:0.17.6
docker pull determinedai/determined-master:a7806b5a
docker pull determinedai/determined-master:a7806b5a6670a0c6a2d9126b004384d322930b73
docker pull determinedai/determined-dev:determined-master-a7806b5a
docker pull determinedai/determined-dev:determined-master-a7806b5a6670a0c6a2d9126b004384d322930b73
docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.17.6
docker pull nvcr.io/isv-ngc-partner/determined/determined-master:a7806b5a
docker pull nvcr.io/isv-ngc-partner/determined/determined-master:a7806b5a6670a0c6a2d9126b004384d322930b73