0.17.1

Changelog

a2ac78b chore: bump version: 0.17.1rc3 -> 0.17.1
262a4cc docs: add release notes for 0.17.1 (#3091)
a0fdf9d fix: write cluster_info.json in all non-cmd task types (#3094)
cabc955 chore: bump version: 0.17.1rc2 -> 0.17.1rc3
6c5c02d fix: avoid race on schema cache load (#3081) [DET-6108]
02308ca fix: report progress correctly for searchers configured in epochs (#3084) [DET-6112]
6d2708f chore: bump version: 0.17.1rc1 -> 0.17.1rc2
114f5c9 fix: update harness to handle telemetry being off (#3085)
82fff64 chore: bump version: 0.17.1rc0 -> 0.17.1rc1
f7f5860 chore: add redirect to documents (#3054)
e102ac8 chore: upgrade sphinx version (#3077)
f9ed8da chore: update experiment and checkpoint imports for consistency (#3079)
f05c32b fix: fix an issue in some CLI aliases not working (#3078)
0ea9b0e fix: update helm push command to helm cm-push (#3075)
a29efdf chore: bump version: 0.17.1.dev0 -> 0.17.1rc0
9b977b9 chore: lock api state for backward compatibility check
6689e62 fix: mispelling [DET-6095] (#3073)
8419fd1 chore: remove flaky tests (#3069)
0e616f2 chore: speed up cli startup time (#3061)
f536460 feat: add Notes tab on experiment pages [DET-4691] (#3048)
cadb2f6 test: stop trying to close modal twice in a row (#3067)
c0e1757 fix: always mkdir default mounted checkpoint_storage host_path. (#3065)
c1aa088 chore: rename cpu containers to aux (#3056)
d8def53 chore: bump version: 0.17.0.dev0 -> 0.17.1.dev0
ff4df83 docs: add release notes for 0.17.0 (#3024)
0fda11f ci: don't depend on badssl.com for test_custom_tls (#3062)
b916abf ci: update gke version (#3051)
693ded3 feat: run db migrations in transactions [DET-5987] (#3025)
6b4ff18 chore: environment bump analytics-python (#3057)
abb3250 feat: add segment tracking python package to harness (#3053)
bda42c5 test: update experiment row kill to handle modal confirmation (#3055)
94fdc50 chore: remove deprecated io-ts any type (#3045)
51453e4 chore: tweak samples_per_second metric to represent all workers (#3050)
68d9aca docs: reorganize the document structure (#3034)
ca594d0 fix: gracefully handle prestart agent failures (#3049)
2a956a1 chore: remove NativeContext and simplify Context inheritance (#3044)
cc26061 Revert "test mmdetection on p3.8xlarge"
1a3036d test mmdetection on p3.8xlarge
e0129f5 fix: Make agent names unique for det deploy local agent-up (#3038)
f0273d1 chore: add server-side portion of external session handling (#3016)
3b1df0c feat: introduce ClusterInfo API (#2946)
85aabd3 chore: added confirmation modal to task kill [DET-6049] (#3035)
ac31bac chore: adding markdown component (#3033)
b21d88e fix: use str for FileLock (#3036)
50eb8f7 chore: rewrite schemas package without typing internals (#3029)
12f7427 feat: cross-compile for powerpc64 (#2828)
9012a78 chore: prefer https to ssh for git dep (#3028)
5058c60 chore: update release note guidelines (#3027)
aa0252f fix: fix nested hparams with grid (#3021)
7b9fd71 test: fix flake from race in idle watcher tests (#3008)
e0e84bf chore: mark open allocs as closed on restart (#3019)
b7f1c3c chore: upgrade to Go 1.17 (#3015)
87e791f chore: lower e2e-webui resource class (#2935)
7f40dab fix: propagate podspec to gc (#3012)
b8afaa4 include load fast flag in 2.6 (#3007)
1898672 feat: allow kubernetes to use priority from exp config (#2956)
6fc5e7e refactor: move app queries and migrations out of internal/db/postgres. (#3014)
2709f56 chore: handle query param jwt for external auth (#2992)
22fcb05 chore: clean notebook readme (#3013)
4d1734c fix: notebook idle check use master port, cert [DET-6013] (#3010)
54e3e56 chore: menu items require keys for newer antd versions (4.x+) (#3011)
3dfcc11 chore: update package json [DET-5846] (#2982)
e5caf86 chore: recover agent websocket flakes [DET-5935] (#2991)
888eb5d feat: add CPU images for TF 2.5 and 2.6 [DET-5877] (#2981)
a388617 chore: minor copy fix (#3009)
44ebc98 fix: save task end times (#3006) [DET-6028]
cf68f17 chore: nit command exit message placeholder (#3003)
d3595e3 chore: add a few timings metrics when sync_required == True (#2996)
a010785 fix: Inline editor for experiment description truncates placeholder text [DET-6024]
70a089b fix: kill trial should send kill (#3001) [DET-6026]
8d7a451 fix: det t describe --metrics API and rendering [DET-6025] (#3002)
76c17f3 fix: bug in notebook README (#3004)
9f4e270 fix: remove refs to workload start_time from det e describe (#2995)
32c41e1 fix: notebook wait page updates for API changes (#2998)
c9b16e5 fix: rename trial job type to experiment (#2997)
e1f7925 chore: fix nil deref in idle timeout watcher (#2993)
24043e6 fix: always helm push latest version (#2994)

Docker images

• docker pull determinedai/determined-master:latest
• docker pull determinedai/determined-master:0.17.1
• docker pull determinedai/determined-master:a2ac78ba
• docker pull determinedai/determined-master:a2ac78ba1ecf397a2a156c9b9b3ed3bee057899d
• docker pull determinedai/determined-dev:determined-master-a2ac78ba
• docker pull determinedai/determined-dev:determined-master-a2ac78ba1ecf397a2a156c9b9b3ed3bee057899d
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.17.1
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:a2ac78ba
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:a2ac78ba1ecf397a2a156c9b9b3ed3bee057899d

0.17.0

Changelog

7e6721b chore: bump version: 0.17.0rc4 -> 0.17.0
954f297 docs: add release notes for 0.17.0 (#3024)
c747aa8 chore: bump version: 0.17.0rc3 -> 0.17.0rc4
b12cb24 chore: mark open allocs as closed on restart (#3019)
aebd71e chore: clean notebook readme (#3013)
2b5be1e fix: propagate podspec to gc (#3012)
b061382 feat: allow kubernetes to use priority from exp config (#2956)
2bf156e chore: bump version: 0.17.0rc2 -> 0.17.0rc3
9f0d23a fix: notebook idle check use master port, cert [DET-6013] (#3010)
c47ae17 chore: bump version: 0.17.0rc1 -> 0.17.0rc2
826f847 fix: save task end times (#3006) [DET-6028]
ed107fd chore: nit command exit message placeholder (#3003)
92d63f6 chore: add a few timings metrics when sync_required == True (#2996)
701458e fix: remove refs to workload start_time from det e describe (#2995)
79f4cd8 chore: recover agent websocket flakes [DET-5935] (#2991)
53beb62 feat: add CPU images for TF 2.5 and 2.6 [DET-5877] (#2981)
527dd21 chore: bump version: 0.17.0rc0 -> 0.17.0rc1
598d7f4 fix: notebook wait page updates for API changes (#2998)
be56db2 chore: fix nil deref in idle timeout watcher (#2993)
3930ced fix: det t describe --metrics API and rendering [DET-6025] (#3002)
7dcd4f6 fix: kill trial should send kill (#3001) [DET-6026]
f0cf24d fix: bug in notebook README (#3004)
3d878eb fix: always helm push latest version (#2994)
1a31a2b chore: bump version: 0.17.0.dev0 -> 0.17.0rc0
070dd4b chore: lock api state for backward compatibility check
0b65b7b feat: det deploy local: remove support for --auto-bind-mount [DET-5948] (#2932)
b15b689 fix: tell mypy to ignore azure (#2990)
bcd959c fix: update cuda for fake tests (#2983)
a0db48e Add support for float16 serialization (#2915)
d482f25 fix: address CVEs in agent & master docker images. (#2989)
66b55b7 chore: update notebook README [DET-6001] (#2985)
9726280 fix: implement boto3 wrapper to allow refreshable credentials [DET-5690] (#2957)
bea9341 chore: StorageManagers operate on uuids, not checkpoint manifests (#2970)
5542a0a chore: confirm with users when running det deploy aws down [DET-6000] (#2984)
8c37d5b chore: update task log response shape (#2986)
576b51f feat: support configuring working directory for tasks [DET-5009] (#2773)
1825c54 fix: make PIDServer send SIGKILL after waiting on SIGTERM (#2976)
64bf390 chore: unify task types [DET-5950, DET-5955] (#2938)
8d98692 chore: popout new tab when clicking on task list links [DET-5998] (#2979)
100f9b1 feat: allow experiment owner to delete their own experiments [DET-5989] (#2977)
36ba4c3 chore: use mock library in doc building (#2968)
086f0db chore: remove -r option since default macos ln doesn't support it (#2971)
5a43b3b chore: remove start_time from get_checkpoints_for_trial (#2975)
8f22270 feat: remove start_time from all workload types [DET-5979] (#2912)
483a24b chore: add STEP_WITH_OPTIMIZER setting for lr scheduler (#2960)
fbb3294 chore: rework tensorboard and checkpoint gc paths (#2948)
2c4f97e chore: fix returning nil error (#2972)
557ea3a chore: update GET raw allocation to account for loss of workload information [DET-5973] (#2911)
722b89f fix: model-hub mmdetection logging (#2964)
28cea50 chore: restore saml auth file to match ee version (#2967)
451f587 refactor: move ee to oss [DET-5937] (#2963)
0eb1390 fix: update logic on when query url should be overridden (#2942)
c4e9743 chore: add support for batch delete of experiments [DET-5224] (#2958)
1f4315d chore: remove trial details start time related stats boxes [DET-5956] (#2944)
e2319f8 chore: rename download model button (#2962)
7344ce1 feat: add detectron2 example (#2918)
a5d0e7d chore: rewrite primary resource allocation query over public.allocations [DET-5972] (#2910)
f070ab2 chore: add warnings on resource manager exits (#2903)
b0b5427 chore: add documentation for model-hub mmdetection [DET-5924] (#2955)
e03bd4d fix: fork nested hp [DET-5945] (#2953)
3b1018c chore: fix a log message (#2945)
5d11b81 fix: extraneous minio warning while using s3 (#2916)
e64befd chore: fix rstrip bug in refresh-ubuntu-amis (#2954)
e0e5912 fix: scroll trial ids with values in trial comparison [DET-5918] (#2933)
560b38e chore: update docs link in notebook webui modal (#2950)
1295bbd feat: add support for nan and infinity metrics [DET-5944] (#2943)
b1e3324 chore: pin mockery version (#2949)
cddff01 fix: make uPlot axis expand to show new data when not zoomed in [DET-5941] (#2928)
5f5b445 chore: add support for throughput profile chart [DET-5596, DET-5732, DET-5913, DET-5923] (#2886)
5d9918d chore: add 1.17 golang build syntax (#2929)
8bc8b25 fix: correct the logic for hiding log preview for completed trials (#2939)
b2da072 feat: trial log preview [DET-5882] (#2871)
9994000 fix: kubernetes link with agent user [DET-5907] (#2927)
bb0fec1 fix: e2e nightly model-hub tests (#2925)
9ddb0ec fix: make clear forbidden vs. unauthenticated [DET-5869] (#2870)
9ded187 ci: replace make -C tools with devcluster. (#2892)
fb08659 ci: unpin pip version, improve py venv cache key. (#2922)
48ee5a1 feat: Support passing an existing EFS to det deploy aws [DET-5737] (#2803)
c8acf89 fix: change styling on "stop experiment" modal [DET-5837] (#2894)
f5dc741 docs: update k8s version to 1.19 >= and <= 1.21 (#2887)
8f2a489 ci: restrict setuptools version. (#2920)
59ca50e chore: fix master/agent Docker image vulnerabilities [DET-5926] (#2914)
7d1c991 feat: add a make devcluster target (#2900)
a6e2172 test: fix batch action misclick on e2e tests [#2872] (#2877)
dd17c3a docs: update idle timeout (#2917)
6508f9d chore: bump version: 0.16.5.dev0 -> 0.17.0.dev0
256a8fd docs: add release notes for 0.16.5 (#2913)
97e5e1f build: remove unneeded build-bindings dependencies (#2898)
ac734f0 feat: support configuring idle timeout for generic commands [DET-4589] (#2787)
190b167 feat: add notebook idle timeout [DET-5517, DET-5519] (#2868)
08cc310 chore: remove old harness profiler (#2901)
1876c04 ci: using_k8s should use det master config -o json. (#2908)
81c23f4 fix: pin version of torchmetrics for docs builds (#2907)
36bf273 chore: fix docstring syntax and remove extra whitespace (#2905)
c0c5bf8 feat: add sync timing toggle for profiling (#2874) [DET-5891]
76f18b1 chore: update master ClusterRole in helm to permit "list" on "events" (#2904)
292f942 fix: det deploy aws --retain-log-group (#2906)
95c859b feat: support mmdetection in model-hub [DET-5471, DET-4558, DET-5609, DET-5610, DET-5474] (#2792)
13fa035 chore: limit overviewstats info height to one line (#2899)
4646dfe fix: include maxval in int hparam range (#2884)
6dc1422 feat: master yaml templates for det deploy aws|gcp [DET-5766] (#2766)
7823b36 fix: enable experiment controls in the header for single trial (#2902)
fc3b83e feat: push architecture (webui side only) (#2855)
eea38c6 feat: push architecture (python side only) (#2771)
00dbd73 feat: push architecture (master side only) (#2776)
5322e9d fix: remove rerender on row selection on Experiment List page (#2897)

Docker images

• docker pull determinedai/determined-master:latest
• docker pull determinedai/determined-master:0.17.0
• docker pull determinedai/determined-master:7e6721ba
• docker pull determinedai/determined-master:7e6721ba8ed9c0ca2d182633982b0a091a6f6d26
• docker pull determinedai/determined-dev:determined-master-7e6721ba
• docker pull determinedai/determined-dev:determined-master-7e6721ba8ed9c0ca2d182633982b0a091a6f6d26
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.17.0
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:7e6721ba
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:7e6721ba8ed9c0ca2d182633982b0a091a6f6d26

0.16.5

Changelog

106b3528 chore: bump version: 0.16.5rc2 -> 0.16.5
af5bd3e docs: add release notes for 0.16.5 (#2913)
2940f8d chore: bump version: 0.16.5rc1 -> 0.16.5rc2
a24a14b fix: pin version of torchmetrics for docs builds (#2907)
c161bce chore: update master ClusterRole in helm to permit "list" on "events" (#2904)
0e5900f feat: add sync timing toggle for profiling (#2874) [DET-5891]
dcd0c41 chore: bump version: 0.16.5rc0 -> 0.16.5rc1
8e4b5d0 fix: enable experiment controls in the header for single trial (#2902)
a036b84 chore: bump version: 0.16.5.dev0 -> 0.16.5rc0
325b29b chore: lock api state for backward compatibility check
863a2d1 feat: expose primitives for pytorch dataloaders (#1937)
beff21b chore: augment fields displayed by describe checkpoint (#2889)
4d08b01 fix: make k8s watchers more resilient [DET-5910] (#2880)
6e4efcc chore: bump gke patch version (#2890)
164560d ci: push latest tag for master and agent Docker images (#2891)
787522e fix: relax imagenet ci target (#2883)
1f6c8cb feat: add det agent disable --drain [DET-5713] (#2827)
70a6f7f fix: profiler metrics without follow should return metrics [DET-5911] (#2879)
053283a build: sunset circleci based react preview (#2881)
920885c docs: update copyright date (#2885)
ed4bedf chore: users can delete their own exps [DET-5901] (#2878)
489b1c6 docs: use virtual environments [DET-5361] (#2862)
f456399 docs: reorganize documentation (#2861)
4ab688c fix: load pre-0.13.8 checkpoints properly (#2876)
1c997da refactor: ban python builtin shadowing. (#2875)
14012f4 fix: don't validate entire expconf on preview-search (#2873)
c8c151b fix: make entrypoint startup-hook.sh eval consistent [DET-5874] (#2847)
e3b5121 chore: bump version: 0.16.4.dev0 -> 0.16.5.dev0
5bab317 docs: add release notes for 0.16.4 (#2865)
dca931e chore: make MetricsBatcherThread safer (#2864)
ad1eb63 fix: uPlot to show values of 0 (#2863)
1d94e50 fix: remove bad switch default (#2859)
9e26e99 fix: dont force nvidia runtime for users using Docker native GPU support (#2854)
174e994 fix: remove visual gap on trial comparison (#2857)
906a76c refactor: apply use settings on other sections (#2849)
f2e4b91 fix: fix row selection and errors on experiement visualization (#2856)
3d89819 fix: test credentials for test_tf_keras_mnist_data_layer_ (#2853)
1bd92e7 style: adjust styles to render exp config and logs to render properly on mobile (#2851)
d0284bb fix: switch container runtime on slot type (#2845)
03cd749 docs: describe how to set task priorities (#2850)
d60f34e fix: reduce imagenet ci time (#2848)
4eb5c3b feat: collect sync_optimizer and backwards pass timings [DET-5724] (#2820)

Docker images

• docker pull determinedai/determined-master:latest
• docker pull determinedai/determined-master:0.16.5
• docker pull determinedai/determined-master:106b3528
• docker pull determinedai/determined-master:106b352802563243c49c52e0a9972e5b04257a25
• docker pull determinedai/determined-dev:determined-master-106b3528
• docker pull determinedai/determined-dev:determined-master-106b352802563243c49c52e0a9972e5b04257a25
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.16.5
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:106b3528
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:106b352802563243c49c52e0a9972e5b04257a25

0.16.4

Changelog

88e26e6 chore: bump version: 0.16.4rc0 -> 0.16.4
69817c4 docs: add release notes for 0.16.4 (#2865)
f87495f docs: describe how to set task priorities (#2850)
093e486 chore: bump version: 0.16.4.dev0 -> 0.16.4rc0
5b6f318 chore: lock api state for backward compatibility check
115461b chore: bump version: 0.16.3.dev0 -> 0.16.4.dev0
1b95dcd docs: add release notes for 0.16.3 (#2774)
b7c8838 feat: support tf 2.6 in TFKerasTrial, update env images. (#2839)
05ddec4 chore: remove some unnecessary conversions and error paths (#2840)
acd4505 fix: support pre-cross_rank versions of horovod (#2841)
e3d76b2 fix: casts non-number HPs to strings (#2837)
892fd66 refactor: responsive table batch [DET-5848] (#2836)
c12cd20 feat: adding profiling metrics to continuous benchmarking (#2796)
b4e0a1a feat: add the ability to set job priorities on the fly [DET-5863] (#2834)
5ca1aac docs: improve Notebook docs (#2811)
94f7b35 fix: reformat porting tutorial (#2833)
1102ced docs: add porting guide (#2624)
6d69c53 feat: add imagenet pytorch example (#2623)
05a488b fix: add in 'just a snapshot' msg to not lose progress on restart (#2830)
cd98af6 chore: clean up priority scheduler code a bit (#2831)
f6c41cf feat: add links to trial pages in trial comparison modal [DET-5850] (#2817)
9cf6389 style: slight style tweaks for trial comparison (#2829)
304b25c feat: allows selection and unselection of hps and metrics in trial comparison table [DET-5854] (#2826)
cbd7408 feat: add ability to compare Trials from Learning Curve and HP Parallel Coords charts [DET-5851] (#2822)
aa43d87 refactor: settings data flow [DET-5625] (#2786)
b937a7b fix: remove unnecessary rows from trial comparison [DET-5857] (#2816)
b7596b7 feat: add Trial ID to header of single-trial experiment page [DET-5816] (#2821)
9648ac1 fix: show tabs independent of trial detail loading [DET-5839] (#2823)
305f138 refactor: allow user to inline edit experiment name [DET-4405] (#2674)
26d395e fix: properly dedup container configs (#2825)
1acad83 fix: nail down installed swag cmd version. (#2824)
ad63487 fix: issue with install and version check of tensorboard (#2819)
6615a3f fix: don't display indefinite spinner if experiment is paused [DET-5849] (#2813)
081b329 fix: size and resize the config monico editor [DET-5736] (#2802)
2107527 chore: upgrade typescript target to es6 (#2812)
24492cf fix: minor trial comparison layout adjustments [DET-5852] [DET-5853] (#2814)
579685b chore: basic python 3.9 support. (#2808)
b4a5872 chore: don't silently drop agent ws failures (#2755)
706b30b fix: correct timing metric chart y-axis label to Seconds (#2810)
46bb3ef chore: increase gRPC recv cap to accomodate equal size shell and experiment context dirs (#2807)
02e3210 update gke version (#2806)
4bfebad feat: add experiment deletion support to the webui [#2752] (#2775)
3486a49 feat: switch to bash in jupyterlab shells. [DET-5791] (#2804)
8112a7e feat: move workloads back to trial overview page [DET-5738] (#2799)
f08ab04 fix: preserve zoom levels between uPlot remounts [DET-5636, DET-5751] (#2797)
9297c84 ci: add label for GKE clusters in CI (#2800)
61886e1 feat: add trial comparison modal [DET-5417] (#2794)
4029250 docs: fix formatting lint failure (#2801)
55d9c6f fix: remove track for constant hps [DET-5815] (#2798)
d941855 refactor: fork and continue trial [DET-5817] (#2765)
fd7cf0d docs: add manual aws modification steps [#2716] (#2749)
8c45251 feat: change k8s preemption scheduler to backfilling scheduler [DET-5398] (#2795)
8931593 test: improve web e2e test stability [#2750] (#2777)
7769d2e ci: add linter for secrets (#2791)
f10108f docs: add git-secrets docs [DET-5830] (#2790)
673cf1d fix: prevent log html injection via unicode [DET-5826] (#2789)
81258ed chore: fix release note lint (#2788)
18a7f5c chore: remove stale install extras from harness python package. (#2767)
23d6a8a fix: only run circleCI step on master branch (#2782)
29cddc0 feat: add --config override to det e create. [DET-5786] (#2769)
e28e871 chore: update rstfmt line length to 100, and reformat all docs. (#2768)
48c5a30 fix: fix indefinite spinner for terminal single-trial experiments (#2772)
28a106c feat: persist checkbox selection when changing page if results are paginated [DET-5416] (#2756)
52c93a5 fix: allow full-sized model definitions through grpc (#2762)
6c913f1 fix: add reload route to allow remounting of same pages [DET-5818] (#2761)
f56255e feat: integrate test results from CircleCI jobs with persistent benchmarking (#2737)
95db1fe fix: fix stale trial data caused by internal react re-route (#2763)
a42b6ad fix: POST /api/v1/experiments/:id/cancel should cancel not kill trials (#2759)
73d5d31 fix: fix wait messaging around undefined trial id (#2757)
f3d6a82 chore: fix log type hyperparameters (#2758)
8db5e26 fix: pull model definitions into containers [DET-5788] (#2753)
866670b feat: async DELETE experiment [DET-5804] (#2741)

Docker images

• docker pull determinedai/determined-master:0.16.4
• docker pull determinedai/determined-master:88e26e66
• docker pull determinedai/determined-master:88e26e66f3da10cb2867bba9e3d3883e51af6c8a
• docker pull determinedai/determined-dev:determined-master-88e26e66
• docker pull determinedai/determined-dev:determined-master-88e26e66f3da10cb2867bba9e3d3883e51af6c8a
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.16.4
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:88e26e66
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:88e26e66f3da10cb2867bba9e3d3883e51af6c8a

0.16.3

Changelog

abc20a3 chore: bump version: 0.16.3rc2 -> 0.16.3
ffff0f4 docs: add release notes for 0.16.3 (#2774)
ef96c9d chore: bump version: 0.16.3rc1 -> 0.16.3rc2
a027c81 fix: allow full-sized model definitions through grpc (#2762)
bfa40d6 fix: add reload route to allow remounting of same pages [DET-5818] (#2761)
28b4121 fix: fix stale trial data caused by internal react re-route (#2763)
e7b37d4 fix: POST /api/v1/experiments/:id/cancel should cancel not kill trials (#2759)
c893777 fix: fix wait messaging around undefined trial id (#2757)
6a55070 chore: bump version: 0.16.3rc0 -> 0.16.3rc1
e226dad chore: fix log type hyperparameters (#2758)
7e38c1b fix: pull model definitions into containers [DET-5788] (#2753)
31656be chore: bump version: 0.16.3.dev0 -> 0.16.3rc0
4ce8500 chore: lock api state for backward compatibility check
d411174 feat: add AKS support [DET-5464] (#2524)
ca44e98 fix: k8s available GPUs indicator [DET-5808] (#2754)
6da6355 feat: add nested hp support [DET-4786] (#2699)
8c5b689 test: set explicit agent slot_type for e2e-tests local and ci env (#2747)
e63e522 fix: correct single trial experiment routing [DET-5789] (#2748)
b4ab5b1 chore: update the jackc/pgx version for ResetSession hook (#2729) [DET-5018]
c0a6d4e fix: hp search configs for darts cnn example (#2725)
9df9099 feat: add filtering by state to trials page [DET-5730] (#2732)
db7437c fix: false positive scaler_state_dict warning (#2745)
bdf86bf refactor: improve fork error message [DET-5795] (#2735)
683fbc6 chore: replace node sass [DET-5806] (#2743)
ff732a1 chore: fix rendezvous address port parsing (#2744)
8cb3540 chore: re-enable automatic updates to the preview cluster [#2680] (#2734)
ab3f2bf chore: add patch to webui proxy (#2742)
22d67ff refactor: clean up task spec 2 (#2698)
a901aa1 chore: Add back Continue Trial button to single-trial experiment page [DET-5749] (#2731)
3deb93f feat: check server reachability on first load [#2739] (#2721)
11a4608 feat: add view logs action to experiment trials [#2714] (#2723)
2e6182e docs: fix some monospace treatment. (#2712)
6a97d46 test: bump test_streaming_observability_metrics_apis timeout (#2738)
7bee752 fix: always terminate streams for terminated trials [DET-5790] (#2728)
d7fd221 fix: restrict google-cloud-storage dependency. (#2736)
90b8700 fix: destroy hp-viz tabs when navigating away [DET-5794] (#2730)
a77e3b9 chore: temporarily disable automatic updates to the preview cluster (#2733)
40a2451 feat: add synchronized query paramaters to several pages [DET-5301] (#2711)
4ea7a1c feat: add descriptive messages to various loading spinners [#2719] (#2718)
fa90df6 feat: enable sorting trials by state [#2673] (#2722)
06628d2 chore: fix a missing loading indicator for task list (#2720)
e84ad9f docs: fixing typo (#2710)
5a825ad chore: fix scientific notation in example yamls (#2688)
953af22 fix: bug in eval for dtrain for question-answering example [DET-5756][DET-5757] (#2707)
32b3050 chore: update model-hub transformers base image [DET-5701] (#2614)
1840ea8 fix: Add legend to Trial metrics graph [DET-5723] (#2663)
ebf0e92 refactor: show "no data" instead of spinner when multi-trial is not available (#2713)
a8e3d8d fix: cpu only preemption [DET-5763] (#2717)
5f4db2d chore: replace proteins_pytorch_geometric example with a better one. (#2706)
837083e chore: bump version: 0.16.2.dev0 -> 0.16.3.dev0
67ade3c docs: add release notes for 0.16.2 (#2709)
c0f5528 docs: EKS Auto-scaling fix [DET-5728] (#2677)
dae469f fix: immediately fetch single trial data when able to [DET-5758] (#2708)
41b73c3 chore: audit and update react dependencies (#2650)
5aad1bc fix: fix an issue with rendering boolean hp values in learning curve table [#2670] (#2694)
65f7ab0 chore: improve dynamic server address support (#2690)
3ce4b51 docs: warn users not to use cpu training with a custom scheduler (#2703)
5f4ff29 revert: "fix: prevents k8s priority scheduler from blocking cpu only training (#2631)" (#2702)
4a52b92 fix: remove proteins_pytorch_geometric from nightlies. (#2701)
d2c1f35 chore: regenerate certificates for multimaster test. (#2704)
3a8a2bc fix: e2e test (#2700)
93e92a9 build: let netlify handle node dependency management (#2693)
d39adbc fix: no data available [DET-5734] (#2695)
488fe33 chore: fix error introduced by rebasing (#2697)
df183b9 feat: allow nightly gpu tests to be requested (#2692)
115e357 chore: clean up task spec (#2662)
c3ef821 fix: preserve subroutes (#2676)
3f99b6c Revert "remove slack user mentions"
f81a1b4 remove slack user mentions
273fc78 chore: move automount path into /run/determined/workdir (#2687)
517b7b1 ci: add preview cluster creation to Circle CI (#2681)
bb4b1cc fix: visual bug in hp plot [DET-5720] (#2689)
4bee18b fix: e2e nightly tests affected by environment upgrades (#2683)
80a56b5 fix: return tensorboard config with GET /api/v1/tensorboards/ (#2685)
c9c1a7f fix: avoid guessing whether loginRedirect is an internal route [#2686] (#2684)
759c9aa feat: Add Efficientdet example (#1733)
e5e0695 fix: correctly interpret minval and maxval for log hps (#2682)
037ead6 build: enable and add Netlify config [#2629] (#2651)
b1d7432 chore: lint det deploy (#2679)
b9340d3 chore: fix a flag in det deploy gcp (#2678)
a7f3da6 feat: improve det deploy [DET-5684] (#2675)

Docker images

• docker pull determinedai/determined-master:0.16.3
• docker pull determinedai/determined-master:abc20a36
• docker pull determinedai/determined-master:abc20a36d08929681fca9e64710ef1189bdbff15
• docker pull determinedai/determined-dev:determined-master-abc20a36
• docker pull determinedai/determined-dev:determined-master-abc20a36d08929681fca9e64710ef1189bdbff15
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.16.3
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:abc20a36
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:abc20a36d08929681fca9e64710ef1189bdbff15

0.16.2

Changelog

2eefec9 chore: bump version: 0.16.2rc6 -> 0.16.2
b7b153d docs: add release notes for 0.16.2 (#2709)
f330072 chore: bump version: 0.16.2rc5 -> 0.16.2rc6
7267a6c fix: immediately fetch single trial data when able to [DET-5758] (#2708)
45fb095 chore: bump version: 0.16.2rc4 -> 0.16.2rc5
27cfbc5 fix: remove proteins_pytorch_geometric from nightlies. (#2701)
c908063 fix: fix an issue with rendering boolean hp values in learning curve table [#2670] (#2694)
cfc2d9a fix: fixes commit 9fec2f9 that cherry picked a revert (#2705)
5d5e3d1 chore: bump version: 0.16.2rc3 -> 0.16.2rc4
ffe8f44 chore: regenerate certificates for multimaster test. (#2704)
dd67fe7 docs: warn users not to use cpu training with a custom scheduler (#2703)
9fec2f9 revert: "fix: prevents k8s priority scheduler from blocking cpu only training (#2631)" (#2702)
8cf0835 fix: no data available [DET-5734] (#2695)
0719b3c chore: bump version: 0.16.2rc2 -> 0.16.2rc3
e226c69 fix: visual bug in hp plot [DET-5720] (#2689)
70a822f chore: move automount path into /run/determined/workdir (#2687)
af7970a fix: e2e nightly tests affected by environment upgrades (#2683)
702328c chore: lint det deploy (#2679)
97e4e60 chore: fix a flag in det deploy gcp (#2678)
6d4397e feat: improve det deploy [DET-5684] (#2675)
d5283f7 chore: bump version: 0.16.2rc1 -> 0.16.2rc2
fd8b813 fix: return tensorboard config with GET /api/v1/tensorboards/ (#2685)
d7627b0 fix: avoid guessing whether loginRedirect is an internal route [#2686] (#2684)
b94d9d2 fix: correctly interpret minval and maxval for log hps (#2682)
7977c2c chore: fix a flag in det deploy gcp (#2678)
f2c37db feat: improve det deploy [DET-5684] (#2675)
ef9b205 chore: bump version: 0.16.2rc0 -> 0.16.2rc1
604f173 chore: bump version: 0.16.2.dev0 -> 0.16.2rc0
09602aa chore: lock api state for backward compatibility check
8f30870 fix: cli auth storage updates on Windows. (#2671)
9aa6758 chore: update images for release (#2672)
e826bfe feat: improve support for pytorch_geometric and custom pytorch batches (#2644)
fc00894 fix/docs: support det-deploy local cluster-up specified directory to bind mount [DET-5431] (#2668)
ae91f6f install tensorboard if not already installed (#2633)
6c65dba fix: pull registry_auth from experiment for tensorboard (#2616)
dc12916 chore: update experiment tag list limits (#2655)
cd0898a chore: simplify resourcetype related states following its removal (#2658)
c8fc329 fix: fix an issue with SPA routing affecting model download [#2648] (#2661)
b9501a8 fix: kube RPs should allow aux tasks [DET-5710] (#2652)
eb57a0e fix: correct for possibility of negative numbers in log range [DET-5717] (#2666)
9996f42 fix: Add page elements to Trial page while data is loading [DET-5718] (#2665)
b969425 fix: Configuration height doesn't go to the full page height [DET-5721] (#2664)
c31f7f4 feat: support auto bind mount as part of det deploy local [DET-5432] (#2610)
c290ab2 fix: improve trial chart tooltip [DET-5712] (#2653)
4f78336 fix: header progress bar [DET-5709] (#2645)
31c87b4 feat: change hyperparameter tab on multi-trial experiment trial pages [DET-5413] (#2642)
cbf6a63 chore: add a readme for react/scripts (#2641)
6bfade1 build: remove java dependency for building webui [#2591] (#2581)
7397a9d chore: make experiment description editable via webui [DET-4398] (#2634)
2e4c995 feat: make det deploy gcp clusters log to GCP Cloud Logging (#2639)
98d7a3a chore: update webui experiment layout for multi-trial and single-trial [DET-5407] (#2595)
15f0f0a fix: make "Continue Trial" button work again [DET-5704] (#2636)
89de31c chore: bump version: 0.16.1.dev0 -> 0.16.2.dev0
40df010 docs: add release notes for 0.16.1 (#2632)
5115495 fix: prevents k8s priority scheduler from blocking cpu only training (#2631)
6454331 fix:nightly tests broken after environment upgrades (#2628)
b21a57f chore: update ZMQ logic in DistributedContext (#2593)
1f945af feat: add rendezvous API (#2420) [DET-5428]
10b3a5c fix: prevent impossible slot requests for notebooks [DET-5690] (#2625)
4b3278c chore: improve error handling for react preview proxy [gh-2622] (#2621)
4f0fa2c chore: Update environments to have a minimal base layer (#2627)
a1b5e4a docs: correct pod spec (#2618)
d612a97 fix: improve det deploy aws messaging for inconsistent stack states [DET-5695] (#2617)
0bd5165 fix: make the gRPC gateway never use a proxy [DET-5689] (#2620)
0b3920e feat: show a loading state while fetching profiler metrics [gh-2605] (#2606)
5f98e0b chore: add tooling to document echo-based apis (#2529)
94246b4 check go mod tidy causes no changes for /proto (#2619)
b09ed81 feat: add API to query an experiments best searcher validation (#2422) [DET-5212]
6acf0d1 chore: update default environment images before release (#2615)
2ab89ed fix: update custom-env docs with new versions (#2611)
d3b55ed fix: linting error in circle ci (#2613)
252351a feat: display active table filter count and option to reset [gh-2584] (#2603)
7d094de fix: print out errors when agent setup script generating fails (#2612)
27e321f fix: avoid presenting gpu uuids in profiler if there are no uuids [gh-2604] (#2608)
09d18c0 fix: helm chart version in Makefile (#2609)

Docker images

• docker pull determinedai/determined-master:0.16.2
• docker pull determinedai/determined-master:2eefec98
• docker pull determinedai/determined-master:2eefec98a0a49856c4a44a91fc5031323d1e04ca
• docker pull determinedai/determined-dev:determined-master-2eefec98
• docker pull determinedai/determined-dev:determined-master-2eefec98a0a49856c4a44a91fc5031323d1e04ca
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.16.2
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:2eefec98
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:2eefec98a0a49856c4a44a91fc5031323d1e04ca

0.16.1

Changelog

de111d2 chore: bump version: 0.16.1rc4 -> 0.16.1
a160523 docs: add release notes for 0.16.1 (#2632)
0f565e2 chore: bump version: 0.16.1rc3 -> 0.16.1rc4
058290e feat: display active table filter count and option to reset [gh-2584] (#2603)
7e52ba1 fix: make the gRPC gateway never use a proxy [DET-5689] (#2620)
c2d9ac8 chore: bump version: 0.16.1rc2 -> 0.16.1rc3
a028858 chore: update default environment images before release (#2615)
16af14f chore: bump version: 0.16.1rc1 -> 0.16.1rc2
aa70a3a fix: print out errors when agent setup script generating fails (#2612)
074bc4a chore: bump version: 0.16.1rc0 -> 0.16.1rc1
53c11d0 fix: helm chart version in Makefile (#2609)
04ec8de chore: bump version: 0.16.1.dev0 -> 0.16.1rc0

Docker images

• docker pull determinedai/determined-master:0.16.1
• docker pull determinedai/determined-master:de111d2e
• docker pull determinedai/determined-master:de111d2e27edee2e6cabcdda2ba01757314caf5e
• docker pull determinedai/determined-dev:determined-master-de111d2e
• docker pull determinedai/determined-dev:determined-master-de111d2e27edee2e6cabcdda2ba01757314caf5e
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.16.1
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:de111d2e
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:de111d2e27edee2e6cabcdda2ba01757314caf5e

0.16.0

Changelog

f5a590b chore: bump version: 0.16.0rc4 -> 0.16.0
091e039 docs: add release notes for 0.16.0 (#2575)
580e60b chore: bump version: 0.16.0rc3 -> 0.16.0rc4
be47a79 docs: update the JupyterLab bump release note (#2567)
e532761 fix: don't return dupes from det model list-versions (#2564) [DET-5640, DET-4248]
5fa3f22 chore: bump version: 0.16.0rc2 -> 0.16.0rc3
438b112 perf: optimizations to query batching fetch profiler metrics [DET-5637] (#2559)
11fff5b chore: bump version: 0.16.0rc1 -> 0.16.0rc2
ffe65cd fix: Change wording on modals that edit configs. (#2562)
89649c7 fix: set elastic ip domain to vpc in det deploy aws (#2557)
e48cd1d fix: dedup BindMounts and Devices on merge (#2560)
9938be7 fix: use model instead of schema struct for de-duping (#2545)
e0c8dec docs: extend docs for the client module (#2556)
2687f4d docs: add python sdk docs (#2547)
5977ce0 chore: also set cli_cert in dtrain worker processes (#2555)
d6edb9e chore: bump version: 0.16.0rc0 -> 0.16.0rc1
62e99c0 chore: fix typos (#2554)
2736600 chore: rename profiler tab in webui (#2551)
4cee9fa fix: Incorrect help link when profiles aren't enabled for a trial. [DET-5621] (#2549)
d623f5f chore: rename start_on_batch to begin_on_batch everywhere (#2553)
61be955 chore: revamp experiment and trial pages header [DET-5406] (#2456)
ee72cdc fix: add bumpenvs for tf-2.5 images. (#2552)
6010642 chore: bump version: 0.16.0.dev0 -> 0.16.0rc0
b910703 chore: lock api state for backward compatibility check
95f7d88 chore: bump version: 0.15.6.dev0 -> 0.16.0.dev0
d5145fe docs: Release notes for 0.15.6. (#2493)
068bb33 fix: prevent zoom reset if chart is already zoomed [DET-5514] (#2525)
3f44c83 fix: stop parsing notebook config on every edit [DET-5605] (#2528)
03b28be chore: fix client for new password handling (#2546)
fe05b0b chore: avoid defaulting to filter by current user [DET-5602] (#2540)
1e945af feat: expose a default Determined in det.experimental.client (#2532)
76230f8 chore: remove swagger-generated python code (#2541)
c7ac21d fix: password handling in python sdk. (#2543)
56dd19d feat: pull tensorboard images from experiment configs (#2544)
48ceaf2 fix: fix hparam string representation failure [DET-5616] (#2539)
8dfa088 feat: pull tensorboard images from experiment configs (#2534)
0ebeba3 chore: fix dropped cert argument in Authentication (#2542)
d0adc51 feat: multimaster Authentication objects [DET-5308] (#2531)
f1c9b1f feat: bump JupyterLab to 3.0.16 [DET-4872] (#2526)
12a8cae chore: bump default environment CPU and GPU images to tf-2.4 (#2523)
caf61c9 docs: add release notes for profiling features [DET-5351] (#2535)
deb4cbf chore: initialize cli_cert in e2e tests (#2530)
81eefc7 chore: bump transformers version for model-hub (#2522)
e9f5947 fix: add init_invalid_hp to master [DET-5569] (#2478)
ccdcaa8 chore: allow non-singleton Authentication (#2513)
0a887e9 fix: trial profiling system metric chart ignoring zero [DET-5505] (#2515)
0d9a540 fix: allow bumpenvs to update nvcr images in helm charts (#2520)
ec89928 feat: provide tensorflow 2.5 image [DET-5522] (#2517)
55c3353 docs: recommend users upgrade to 0.16.0 to avoid k8s master crashes (#2518)
a2f6fc2 chore: improved pynvml usage by profiler [DET-5394] (#2487)
a06d3a2 chore: minor edits to cli behaviors (#2519)
2316057 fix: add back bindmounts entry to command's default config (#2521)
3d34e1c fix: notebook modal improvements [DET-5599] (#2511)
6db8263 feat: add experiment notes & name [DET-5352] (#2307)
17976404 chore: update urllib3 (#2504)
49aec0d feat: support back-filling in the priority scheduler [DET-5397] (#2436)
aec1074 chore: handle error when loading notebook config (#2512)
09fca00 feat: add bind mounts to task container defaults [DET-5362] (#2516)
4068fde chore: collect prometheus metrics (#2501)
ed896c7 fix: python api create experiment bug (#2510)
0c9ec27 fix: avoid rc dev release mismatch notifications (#2405)
4fd3326 chore: task list filters [DET-5390] (#2466)
1f49553 test: add e2e tests for profiling features [DET-5245] (#2481)
ec9932d chore: upgrade ws to patch security vulnerability (#2505)
3857f94 chore: add experiment name to breadcrumb on trial detail page [DET-5284] (#2318)
87b1e59 docs: add release note for printable config (#2507)
212aa93 chore: disable profiling after restart [DET-5424] (#2486)
24432fe docs: add profiling how-to [DET-5209] (#2384)
2b04bf0 chore: fix TrialsSnapshotResponse comment typo (#2492)
1ca42b4 chore: fix TF version detection and RNG usage in test (#2500)
106294a chore: migrate away from spot checks and move towards waiting for an expected case (#2495)
eecc446 fix: generating printable master config does not alter original (#2502)
bf9b3ac fix: observability webui fixes [DET-5567][DET-5246][DET-5506][DET-5531][DET-5530][DET-5571] (#2488)
5b73278 chore: improve profiler throughput collectors (#2490)
55b122e chore: remove native init() functions [DET-5574] (#2480)
6ac0268 chore: add testing for eventually schema [DET-5560] (#2467)
6f86594 chore: remove trial old messages and consolidate others (#2464)
bae9c2d chore: fix some semi-broken unit tests (#2483)
3f9f2da fix: ship gpu_free_memory correctly [DET-5508] (#2497)
0dae801 chore: add non-streaming APIs for trial profiler endpoints (#2484)
0b0e9ca chore: update eslint-no-unused-vars to handle special cases (#2496)
d81f8ad fix: notebook modal bugs [DET-5573] (#2476)
8ee598d chore: improve performance of tfevent file filtering (#2469)
341fb4f chore: trim unused parts of rendezvous info (#2381)
ba07a04 chore: promote profiler APIs out of unimplemented (#2485) [DET-5587]
3f53289 fix: send all batches from harness profiler [DET-5566] (#2473)
c520187 chore: deprecate det.experimental.create_trial_instance() (#2479)
b0f57d6 fix: ProfilingAgent serializing timestamps incorrectly (#2482)
6a67383 fix: propagate slots when it is 0 (#2477)
4b97010 chore: measure profiler timings with time.time() (#2475)
2e38dfa chore: reword README for schemas (#2474)
3d6e73d fix: show x axis label on all plots [DET-5500] (#2471)
2e83f22 fix: make tf estimator dtrain work with tf 2.5 [DET-5563, DET-3762] (#2468)
f893eee fix: timing metric chart x-axis tick off [DET-5501] (#2472)
aa8d442 chore: log running of migrations (#2463)
36139a1 docs: add instructions to use dtrain workflow for inference with PyTorch (#2386)
66c6452 feat: hook ProfilerAgent into harness and add profiler timings [DET-5062, DET-5204] (#2348)
c52c616 chore: move run increment to allocation not termination [DET-5559, DET-5450] (#2462)
feac8cf feat: add launch notebook modal [DET-5376] [DET-5377] [DET-5380] [DET-5378] [DET-5379] [DET-5375] (#2398)
7c17856 chore: catch ruamel.yaml Duplicate Key Errors and format for users [DET-5542] (#2450)
2584c5b chore: rem to px [DET-5327] (#2433)
ddf8693 fix: allow custom registries with determined env images [DET-5556] (#2465)
8c1d0a9 fix: cleanup iter(DataLoader) before exiting [DET-5558] [DET-5554] (#2459)
2c3bfa3 fix: use user preferences when no search params are present (#2460)
80f4375 chore: disable dashboard recent tasks tests temporarily (#2461)
7f1c61d feat: det deploy --image-repo-prefix for pulling images from a custom docker repo (#2454)
a517040 fix: synchronize pods actor startup in k8s resource manager [DET-5536] (#2453)
ea4566f fix: update Buf image and CLI usage (#2455)
8092072 chore: bump buf and protoc version [DET-5534] (#2446)
92bf2c6 fix: prevent concurrent updates to a single expconf object [DET-5543] (#2451)
ea66301 revert added example model (tf classification) (#2452)
71a3502 fix: prevent spot resource pool contention [DET-5349] (#2423)
8def156 cli: small rewording in shell help (#2448)
bac3924 ci: regen buf image with buf 0.12.1 (#2447) [DET-5534]
193ac65 docs: fix broken links (#2439)
da7fe34 fix: introduce LegacyConfig for tensorboard and checkpoint gc [DET-5533] (#2444)
a9f0fe8 fix: omit internal fields in previewed notebook [DET-5523] (#2434)
a690381 fix: allow EOL searchers in configs only [DET-5526] (#2445)

Docker images

• docker pull determinedai/determined-master:0.16.0
• docker pull determinedai/determined-master:f5a590b8
• docker pull determinedai/determined-master:f5a590b8e8b0f589f8086111c93a42f92760041c
• docker pull determinedai/determined-dev:determined-master-f5a590b8
• docker pull determinedai/determined-dev:determined-master-f5a590b8e8b0f589f8086111c93a42f92760041c
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.16.0
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:f5a590b8
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:f5a590b8e8b0f589f8086111c93a42f92760041c

0.15.6

Changelog

0c9ee55 chore: bump version: 0.15.6rc3 -> 0.15.6
2545084 docs: Release notes for 0.15.6. (#2493)
d7f41bc chore: bump version: 0.15.6rc2 -> 0.15.6rc3
bfad801 chore: move run increment to allocation not termination [DET-5559, DET-5450] (#2462)
fa03ce8 ci: regen buf image with buf 0.12.1 (#2447) [DET-5534]
0c348b8 chore: bump version: 0.15.6rc1 -> 0.15.6rc2
6546b65 chore: catch ruamel.yaml Duplicate Key Errors and format for users [DET-5542] (#2450)
4e3ade0 fix: allow custom registries with determined env images [DET-5556] (#2465)
730daeb fix: cleanup iter(DataLoader) before exiting [DET-5558] [DET-5554] (#2459)
d891b90 fix: synchronize pods actor startup in k8s resource manager [DET-5536] (#2453)
c2f08d2 fix: use user preferences when no search params are present (#2460)
ae8d54e revert added example model (tf classification) (#2452)
b898f11 fix: prevent spot resource pool contention [DET-5349] (#2423)
a7ce160 docs: fix broken links (#2439)
14300c4 cli: small rewording in shell help (#2448)
80973c1 chore: bump version: 0.15.6rc0 -> 0.15.6rc1
026a929 fix: allow EOL searchers in configs only [DET-5526] (#2445)
71166d1 fix: introduce LegacyConfig for tensorboard and checkpoint gc [DET-5533] (#2444)
ecf0b66 fix: omit internal fields in previewed notebook [DET-5523] (#2434)
396f492 chore: bump version: 0.15.6.dev0 -> 0.15.6rc0
016c33d chore: lock api state for backward compatibility check
cd5c939 fix: webui observability show chart only if metrics are available [DET-5418] (#2424)
6f67799 docs: notify users of coscheduler behavior [DET-5150] (#2442)
861c19a fix: resource pool not saved in the DB [DET-5485] (#2435)
343d810 chore: whitelist eventually from schema linter
25f6f3a feat: add eventually extension to schema [DET-5520] (#2432)
2f9dcff chore: Prevent _swagger from being formatted by make -C harness fmt (#2440)
4643748 docs: update procedure for latest NVIDIA drivers on GKE (#2429)
4f2a16f chore: minor copy fix for alert box spaces (#2438)
6e255d0 fix: make profiling schema more lenient [DET-5497] (#2409)
8c96541 chore: update OS and other language in Terraform modules [DET-4276] (#2415)
5c4b502 chore: reduce minimum char to fuzzy match for omnibar (#2430)
fe71847 fix: correct lint issues (#2437)
c7dcd41 chore: more eslint rules [DET-5513] (#2426)
d272e89 fix: observability webui widen dropdowns so the entire string is readable [DET-5503] (#2425)
55ff22e fix: fix convergence and distributed tests for tensorflow example (#2431)
c055571 docs: using det shell as a remote shell in IDEs. (#2428)
c71ff13 feat: omnibar initial support [DET-5374] (#2308)
57504fe chore: update default images (#2427)
4f7acab fix: merge logic for union-type configs [DET-5486] (#2410)
366b82b feat: det shell option to show ssh command for use in IDE [DET-5462] (#2407)
98b5a1e chore: table head style update (#2419)
e938ad1 chore: terminate /api/v1/trials/:id/avialable_series on trial termination (#2418) [DET-5499]
e3f4fb9 fix: improve tqdm rendering in the web ui (#2320)
2bf8534 Disable tests that will never pass on mac os x (#2417)
077bec8 docs: resource pool fixes (#2408)
8da96e9 fixed typo in custom custom docker configuration (#2413)
d7aa85f docs: update create_experiment (#2416)
0719c1d fix: fix an issue with parsing old exp config labels [DET-5487] (#2411)
6b260e7 chore: only return the port binding appropriate for the proxy [DET-5495] (#2401)
f9e099e docs: update parameter string (#2412)
d075392 feat: python-sdk [DET-5371] (#2317)
74b4e25 feat: add new multiclass text classification example for tensorflow [DET-5277] (#2396)
8e41491 feat: support more types of CPU instances on AWS [DET-4939] (#1907)
58b7e02 feat: experiment list search [DET-5460] (#2392)
66158f0 chore: update trial page overview layout [DET-5411] (#2389)
1d7f049 chore: upgrade timeago-react for react v17 (#2404)
ca4ab7a fix: correct resource pool pagination and make sort sticky [DET-5482] (#2403)
c0e23f4 chore: add zmq-based IPC to the DistributedContext (#2373)
3a625b4 chore: make pip happy again (#2399)
3b08403 fix: add max size limit metrics [DET-4878] [DET-4783] (#2387)
212dd09 chore: remove upstreamed gradient aggregation test (#2406)
8287dbe fix: correct the url search param setting for archived (#2402)
41aea03 docs: Release notes for 0.15.5. (#2397)
a6e3d80 chore: bump version: 0.15.5.dev0 -> 0.15.6.dev0
950a591 docs: deprecate old master configuration fields (#2395)
c0821a9 fix: wire up support for plain-string image config (#2393)
09e1dd8 ci: temporarily remove flaky tests (#2394)
ea3f932 chore: Edit docs for typos (#2391)
9918826 feat: support push metric APIs internally [DET-5215] (#2315)
7b62b33 chore: widen the trial link on experiment detail page [DET-5459] (#2390)
c8dd9fd feat: add SlotsPerAgent in resource pool API (#2383)
993a526 fix: Fix nightly gpu tests for pytorch word language model [DET-5226] (#2388)
fb60a58 chore: move trial logs in a trial detail page tab [DET-5410] (#2365)
0f3ba70 refactor: experiment list native filters [DET-5389] (#2378)
7b58a2d chore: move Trial Information table in a dedicated Trial page tab [DET-5434 (#2372)
e1e4b9e feat: Add PyTorch Word language Modeling example to Determined's Example [DET-5226] (#2352)
9e20c3c chore: unrevert and fix "actually use expconf in the master" (#2382)
8560a96 chore: remove unused protobuf imports (#2336)
f41e788 chore: update gke version (#2385)
1b134e3 chore: simplify tensorboard request msg (#2377)
30ab146 chore: close agents on websocket closures (#2380)
7d1509b docs: spelling fixes in model hub (#2379)
d0c6651 fix: det deploy gcp support for terraform 0.15 [DET-5449] (#2376)
cde5700 chore: revert "actually use expconf in the master" (#2375)
108462f chore: move trial hyperparameters in a dedicated trial page tab [DET-5412] (#2364)
3197cc3 build: remove webui and docs as direct master dependencies (#2363)
3a545cd feat: add a preview parameter to the notebook launch API (#2359)
fd145c9 chore: remove redundant model_hub line from bumpversion. (#2374)
1f573da docs: Release notes for 0.15.4. (#2370)
b7d3f2d chore: bump version: 0.15.4.dev0 -> 0.15.5.dev0
571f321 chore: actually use expconf in the master [DET-4885] [DET-4009] (#2310)
cd86fa2 fix: fix task pagination filters not taking effect [DET-5442] (#2367)

Docker images

• docker pull determinedai/determined-master:0.15.6
• docker pull determinedai/determined-master:0c9ee55c
• docker pull determinedai/determined-master:0c9ee55c459e6407e0df60cf5db2805dc38865c5
• docker pull determinedai/determined-dev:determined-master-0c9ee55c
• docker pull determinedai/determined-dev:determined-master-0c9ee55c459e6407e0df60cf5db2805dc38865c5
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.15.6
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0c9ee55c
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0c9ee55c459e6407e0df60cf5db2805dc38865c5

0.15.5

Changelog

5fe959f chore: bump version: 0.15.5rc1 -> 0.15.5
eb1d821 docs: Release notes for 0.15.5. (#2397)
40640bc chore: bump version: 0.15.5rc0 -> 0.15.5rc1
a0f8ae6 docs: deprecate old master configuration fields (#2395)
7019b17 chore: bump version: 0.15.5.dev0 -> 0.15.5rc0
6c51726 chore: close agents on websocket closures (#2380)
124f04c chore: bump version: 0.15.4 -> 0.15.5.dev0

Docker images

• docker pull determinedai/determined-master:0.15.5
• docker pull determinedai/determined-master:5fe959f6
• docker pull determinedai/determined-master:5fe959f61237b90b6af68999440fe6f52f734492
• docker pull determinedai/determined-dev:determined-master-5fe959f6
• docker pull determinedai/determined-dev:determined-master-5fe959f61237b90b6af68999440fe6f52f734492
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:0.15.5
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:5fe959f6
• docker pull nvcr.io/isv-ngc-partner/determined/determined-master:5fe959f61237b90b6af68999440fe6f52f734492