https://github.com/kubeflow/katib

sort by:
Revision Author Date Message Commit Date
8986a7f handle the GetEvaluationResult() return None problem 05 March 2019, 23:48:55 UTC
2f415f3 fix a potential problem 05 March 2019, 21:46:39 UTC
d328904 fix several problems 05 March 2019, 04:49:21 UTC
ef3ff96 language improvement in README.md 02 March 2019, 08:15:20 UTC
deba168 adjust To Do 02 March 2019, 07:49:49 UTC
e6234d1 supoort multiple trials 02 March 2019, 04:05:15 UTC
4d8c599 Add support for parallel studyjobs (#404) * Add support for parallel studyjobs * fix a typo * Reorganize the program a little bit * fix a typo * fix a typo 27 February 2019, 03:05:45 UTC
87a31f3 Add separable/depthwise convolution, data augmentation and multiple GPU support (#393) * add separable/depthwise convolution in operation library * add ENAS example StudyJob yaml * remove ENAS example, add data augmentation, add multiple GPU support 27 February 2019, 01:33:48 UTC
4d031e7 Add create time to Trial API (#410) * Add create time to Trial API * Add Trial create time information * Fix UT for db 27 February 2019, 00:32:45 UTC
26da3ea Metric collector must fail on error (#405) * Fail when unable to collect logs * Set backlimit to 0 for jobs 26 February 2019, 04:32:34 UTC
6b75138 add latest tag for katib images (#409) 25 February 2019, 17:37:16 UTC
46d2dc7 add build and test for suggestion nasrl (#401) 22 February 2019, 02:35:03 UTC
d6a67ea Database APIs for NAS updated (#394) * FINAL PUSH * FIX TESTS * new lock * new lock * small fi * DELET SPACE * deleted ununsed function 21 February 2019, 01:37:09 UTC
3bb8b54 Suggestion for Neural Architecture Search with Reinforcement Learning (#339) * Suggestion for Neural Architecture Search with Reinforcement Learning * Add NAS RL Suggestion * Fix new line * set json format for GetSuggestion() * finish trial return in GetSuggestion(), finish GetEvaluationHistory, and fix bugs * fix a bug in GetEvaluationResult() * fix bigs in GetEvaluationResult * fix an error in GetEvaluatinResult * Add python Katib api * Remove unnecessary requirements * add about for suggestion * rename to README * Add picture explanations; make the printouts more organized * fix typos * fix some small problems * Fix several problems * Fix a typo * fix some problems * small fixes * Suggestion do not need to handle uncompleted trials * fix a small problem 21 February 2019, 00:53:59 UTC
5a1a791 add validating webhook for studyJob (#383) * add validating webhook for studyJob If create/update a studyJob with bad CR manifest or invalid configuration, k8s api server will reject the request. Fixes: #314 * add test * allow check "kubectl" error code 20 February 2019, 17:40:23 UTC
8a89b9e Removing Operator specific handling during a StudyJob run (#387) * Removing Operator specific handling during a StudyJob run * Return empty in error 20 February 2019, 06:19:50 UTC
edecd39 Delete modeldb from unit tests (#391) * Delete modeldb from unit tests * Add library to interface test 20 February 2019, 00:41:30 UTC
c0f2f07 show studyjob condition when run kubectl get (#389) 19 February 2019, 03:21:42 UTC
ee62c33 Training Container with Model Constructor for cifar10 (#345) * Training Container with Model Constructor for cifar10 * fix a small bug * make num_epochs a parameter 15 February 2019, 02:23:48 UTC
3706fce add studyjob python client (#379) 14 February 2019, 18:15:03 UTC
1de9307 fix wrong example (#378) 14 February 2019, 18:14:53 UTC
03ca08f Upgrading and controller runtime k8s to 1.11.2 (#376) 14 February 2019, 16:36:02 UTC
a5c8e02 Properly initialize CI cluster credential (#360) It has been using the cluster where argo ran 14 February 2019, 05:32:03 UTC
41a5a2e Include go dependencies in developer-guide.md (#369) Looks like Google protobufs might also be a dependency? 13 February 2019, 19:27:47 UTC
d6ea2d5 fix invalid memory address (#368) 12 February 2019, 02:47:44 UTC
421cbff Fix presubmits (#363) * Fix typo * Fix gcloud builds submit command * Use printf() instead of print() 08 February 2019, 06:07:13 UTC
0ea34b1 Katib 2019 Roadmap (#348) * roadmap * Fixing format * Add links to github issues * Fix comments 01 February 2019, 04:28:57 UTC
afee0c3 Update OWNERS (#350) 29 January 2019, 05:43:25 UTC
f11c13e Extend Katib API for NAS jobs (#327) * Add fields to studyjob structure * Change nasjob yaml file * Change parameter type * Add Parameter Type=range * Change API * Change input size * Reset API structure * Change StudyJob API structure * Remove Range parameter * Fix api.proto * Fix gopkg.toml * Remove old nasjob file * Fix nasjob.yaml * Add custom suggestion * Add blank NAS suggestion Change Katib API to process yaml file for NAS * Add correct YAML file for NAS example * Fix newline * Change StudyID to 1 * Add jobType parameter in Parsing * Remove changes in manager * Add NasConfig inside Yaml file * Fix name in nasConfig * Fix get StudyConfig in NAS * Add JobType in all services * Add job_type in bayesian_service * Add pointers in NasConfig structure * Fix Pointer in API * Add consts for jobType Remove return from populateCommonConfigFields * Move const jobType to const file * Remove Range parameter * Modify YAML file for NAS jobs * Add getStudyJobType function in GRPC server * Add blank GetStudyJobType func in manager * Fix metrics collector * Remove jobType from getStudy * Remove getStudyJobType from manager * Add NAS RL yaml deployment * Change worker to GPU * Clean nasrl suggestion * Add -u inside training-container * Fix namespace in worker template 29 January 2019, 01:23:06 UTC
f4026e4 ignore tfjob/pytorch job if corresponding CRD not created (#335) * ignore tfjob/pytorch job if corresponding CRD not created * update log message * only ignore NoMatchError when watch CRD * refactor func name for watch error 25 January 2019, 00:26:22 UTC
c67892f Clarify the example UI is generated by random-example. (#333) 24 January 2019, 16:50:31 UTC
a777721 only try to delete study info in db when in need (#342) 24 January 2019, 01:16:26 UTC
8545970 omit empty fields for studyjob status (#336) 22 January 2019, 19:40:17 UTC
15bbcae Update pytorch example with latest image (#329) * Update pytorch example with latest image * Update pytorch example docker image 18 January 2019, 19:07:12 UTC
a24c428 Fix typo (#330) 18 January 2019, 00:00:00 UTC
d41f8e8 Add information how to run TFjob and Pytorch examples in Katib (#321) * Add doc for tfjob and pytorch examples in Katib * Add contents * Fix README * Fix link to examples in README * Fix README * Add information about Katib UI and status of StudyJob * Add Ambassador information 16 January 2019, 15:30:01 UTC
0ed361c Add xgboost example using Bayesian optimization (#320) * Add xgboost example * Add comments for ames example 15 January 2019, 23:22:00 UTC
4a69776 katib should be able to be deployed in any namespace (#324) 15 January 2019, 02:32:07 UTC
3c37f31 Adding distributed pytorch example for katib (#309) 08 January 2019, 10:56:59 UTC
9aa90fa minor fixes (#307) 08 January 2019, 10:22:02 UTC
f78a108 delete obsolete data in db (#315) * delete obsolete data in db * add delete study test * make sure trials and workers deleted when study deleted in ut test 07 January 2019, 15:22:39 UTC
fae6aa5 add bestTrialId to statusJob status (#312) * add bestTrialId to statusJob status * generate mock and add bestworkerid 03 January 2019, 03:42:45 UTC
f24889c Add api doc (#303) * add api doc Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix typo Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add instructions for update api files and docs Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix typo Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 25 December 2018, 17:22:57 UTC
1295f45 validate studyJob when first reconcile it (#308) * validate studyJob when first reconcile it Fixes: #297 * use 3rd-party uuid instead of self-define one k8s.io/apimachinery/pkg/util/uuid is used in kubernetes source code 23 December 2018, 02:11:11 UTC
cbe91f8 add hougangliu as a reviewer (#310) 22 December 2018, 05:51:27 UTC
9baabbf Adding to OWNERS file (#304) * Adding to OWNERS file * adding to reviewers 21 December 2018, 13:57:09 UTC
b11b81d sync up worker status all the time (#299) Fixes: #298 20 December 2018, 04:53:31 UTC
bca0b58 studyJob with non-kubeflow namespace cannot work (#302) 19 December 2018, 18:02:49 UTC
8e89813 Adding master pod check for default metric collector (#300) 19 December 2018, 15:03:34 UTC
07e0fd2 reduce some redundant code (#296) 19 December 2018, 01:24:56 UTC
28c5b1c Extend studyjob client API (#288) * Add namespace parameter to studyJob client API * Change if statement for namespace * Create func getNamespace 16 December 2018, 15:43:49 UTC
4be865e fix deploy (#284) 16 December 2018, 15:43:43 UTC
eb4a35b update Readme (#295) A trial can be corresponds to a k8s job, TFJob and PyTorchJob now. Not only k8s job any more. 16 December 2018, 15:34:39 UTC
5a7977d fix studyJob status suggestionCount mismatch error (#290) Fixes: #289 14 December 2018, 15:14:46 UTC
41e8f7d fix invalid worker kind issue (#287) * fix invalid worker kind issue studyJob should go to 'Failed' status when worker kind is invalid * add PyTorchJob as valid worker job kind 14 December 2018, 01:18:22 UTC
33b2e58 get metricscollector by API (#292) Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 13 December 2018, 20:00:04 UTC
f16aecc Support Pytorch job in Katib (#283) * Pytorch support in Katib * Adding pytorch worker kind to metrics collector * Updating Gopkg * Adding sleep * Changing the worker name * Adding gcr image 13 December 2018, 16:32:46 UTC
5527e34 Update k8s cluster version to 1.10 (#286) 12 December 2018, 17:01:34 UTC
67eca98 Enrich GUI (#264) * allow to create studyjob from UI Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * show success alert Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add rbac for ui Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix bug Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * rebase master Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add metrics collector manager to UI Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix typo Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 11 December 2018, 07:22:12 UTC
86cddd3 update README (#281) 11 December 2018, 06:46:34 UTC
1c707dc fix typo error for MinikubeDemo (#282) 11 December 2018, 00:30:40 UTC
f8590e0 fix typo error (#280) 10 December 2018, 06:17:24 UTC
edf6cb5 add e2eTest of each suggestion algorithm (#265) * random&grid * hyperband * add hyperband test * add grid case check 09 December 2018, 13:47:06 UTC
f4913b3 Allow studyjobcontroller to delete pods (#278) 09 December 2018, 12:52:52 UTC
c8efb35 Fix katib ui resource paths (#277) 07 December 2018, 16:42:11 UTC
36d8d25 Implement gRPC Health Checking Protocol + add readiness/liveness probes to vizier-core (#270) * Ensure vizier-core never been stuck too long waiting for DB conn Signed-off-by: Koichiro Den <den@valinux.co.jp> * Add standard Health gRPC service Signed-off-by: Koichiro Den <den@valinux.co.jp> * Change db.New to return error instead of exit(1) with log.Fatal Signed-off-by: Koichiro Den <den@valinux.co.jp> * Add SelectOne() to VizierDBInterface Signed-off-by: Koichiro Den <den@valinux.co.jp> * Rename import for later convenience Signed-off-by: Koichiro Den <den@valinux.co.jp> * Implement and register Health Server for Katib manager Signed-off-by: Koichiro Den <den@valinux.co.jp> * Add readiness/liveness probes to vizier-core Signed-off-by: Koichiro Den <den@valinux.co.jp> * Update test codebase Fixes: 61ac5607353 ("Add SelectOne() to VizierDBInterface") Signed-off-by: Koichiro Den <den@valinux.co.jp> 05 December 2018, 09:12:00 UTC
3516dda POC: Katib integration with tf-operator (#267) * TF operator part 1 * Add consts * Fix * Update worker; fix schemes * Change example * Add rbac rules * Add crd * Add sleep for debugging * Log cluster name * Remove unrelated change * use katibapi.State 05 December 2018, 08:33:36 UTC
55f125c fix make timing (#271) 05 December 2018, 07:02:33 UTC
f863b87 Add Update{Study,Trial} (#269) Only tested with unit tests. Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> 05 December 2018, 05:13:30 UTC
0e3e890 add Richard Liu to OWNERS (#274) Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 04 December 2018, 02:57:06 UTC
211c6ba fix uncompleted value in ui (#238) Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 04 December 2018, 01:58:23 UTC
1104524 fix bayesian optimization suggestion (#251) * fix bayse optimization suggestion Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add bayseopt-example Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * reset x_train in burn-in Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * validate parameters Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 04 December 2018, 01:24:06 UTC
72a0fc0 Prevent pod restarts caused by slow db boot (#261) * Add readinessProbe for vizier-db Signed-off-by: Koichiro Den <den@valinux.co.jp> * Fix MYSQL_ROOT_PASSWORD Fixes: 67e94c7697bd ("Set MYSQL_ROOT_PASSWORD via Secret (#253)") Signed-off-by: Koichiro Den <den@valinux.co.jp> * Add simple loop to wait for DB connection successfully opened Signed-off-by: Koichiro Den <den@valinux.co.jp> 30 November 2018, 12:41:56 UTC
3f5462d add UT of each suggestion algorithm (#237) * add random algorithm UT * add grid algorithm UT * add hyperband algorithm UT * fix typo * fix typo * add some tests * change various ParameterType pattern * add gengrid() test * fix significant figure 30 November 2018, 12:06:00 UTC
24160cb Downgrade kubernetes dependency to 1.10.1 (#256) * downgrade to 1.10.1 * Delete pods * Fix job-name * Set successfulJobsHistoryLimit to 0 * Add comments 28 November 2018, 06:52:51 UTC
b7145b3 Fix incorrectly set namespace (#260) Commit b6f8e07d26a ("Update manifests (#246)") has just changed the namespace as a whole. This new manifest should be updated as well. Fixes: 67e94c7697b ("Set MYSQL_ROOT_PASSWORD via Secret (#253)") Signed-off-by: Koichiro Den <den@valinux.co.jp> 26 November 2018, 10:04:51 UTC
67e94c7 Set MYSQL_ROOT_PASSWORD via Secret (#253) * Set randomly generated MYSQL_ROOT_PASSWORD via Secret Signed-off-by: Koichiro Den <den@valinux.co.jp> * Seperate manifest for MYSQL_ROOT_PASSWORD, "test" being set by default Signed-off-by: Koichiro Den <den@valinux.co.jp> * Update run-tests.sh Fixes: 5312459c28f7 ("Set randomly generated MYSQL_ROOT_PASSWORD via Secret") Signed-off-by: Koichiro Den <den@valinux.co.jp> 22 November 2018, 05:59:22 UTC
63dc070 update UI (#255) Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 20 November 2018, 23:57:25 UTC
e5e2dcd Refactor studyjobcontroller (#254) * Refactor studyjob controller * Refactor * Go format files * More refactor * Rename studyjobcontroller to studyjob 20 November 2018, 23:18:57 UTC
597064a Change deploy.sh for Minikube example (#252) * Change deploy for Minikube Example * Change namespace to kubeflow in Minikube example * Delete lines about modeldb from deploy 20 November 2018, 08:19:24 UTC
206bcaa Add mysql based unit tests (#243) Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> 20 November 2018, 01:43:06 UTC
b6f8e07 Update manifests (#246) * change namespace katib -> kubeflow Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * change namespace of tfevent-mc 19 November 2018, 04:58:32 UTC
f7aff4a Add texasmichelle as reviewer (#247) 16 November 2018, 03:05:02 UTC
94b138a Tf event mc (#235) * add tf-event mc Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add tfevent mc ci Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add tfeventmc doc Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add comment and use logger Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 16 November 2018, 01:26:56 UTC
9d59a10 Fix typos for json and objective (#242) Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> 14 November 2018, 06:14:39 UTC
29e53b8 Add richardsliu to OWNERS/reviewer (#239) * Add richardsliu to OWNERS * Add richardsliu as reviewer 13 November 2018, 02:11:21 UTC
a01f482 add starttime and completiontime to worker (#236) 08 November 2018, 08:55:46 UTC
5e51974 Fix typo (#233) * correct "purse" to "parse" * correct "Doubel" to "Double" * Update push-model.go fix lowercase * Update push-study.go use lowercase 05 November 2018, 20:31:38 UTC
04837a4 More DB unit tests (#234) * Fix EarlyStopParam and SuggestionParam DB methods GetEarlyStopParamList and GetSuggestionParamList mixed up the column order and they returned nothing. Also, SetEarlyStopParam didn't return an ID properly. Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> * Add more DB UTs Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> 05 November 2018, 07:47:01 UTC
8e90513 Fix the build script after #208 (#231) Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> 02 November 2018, 05:44:46 UTC
9f87fd8 Only retry an INSERT operation on unique constraint violation (#229) The retry logic is used to generate an unique ID, but if there is another error the DB code can fall into an infinite loop. Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> 01 November 2018, 06:00:42 UTC
0bc5182 New UI for Katib (#208) * add ui Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add ui Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * update test and doc Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix test Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix test Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix test Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * remove modelDB Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix test Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * refactor Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add loading img Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix test Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * Add loading image Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * refactor Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add root redirection Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * add latestLog flag to GetWorkerFullInfo Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix test Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 29 October 2018, 04:20:23 UTC
7eeea12 fix slice range (#226) 28 October 2018, 09:46:16 UTC
13373d2 More db tests (#225) * Remove obsolete comments and an import Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> * Add Worker UTs Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> 25 October 2018, 03:22:52 UTC
106235b Fix storelogs (#222) * Fix StoreWorkerLogs The function has been storing into worker_metrics with duplicates and wrong timestamps for some time. The fix changes the worker_lastlogs DB table definition. DBs must be recreated. Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> * Add foreign key constraints to worker log DB tables and tidy up formatting This patch make sure worker_* rows have matching row in the worker table. Also changes multi-line string formatting for readability. Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> 24 October 2018, 04:00:15 UTC
4dc1aed Check errors in order to avoid SEGV (#219) Signed-off-by: IWAMOTO Toshihiro <iwamoto@valinux.co.jp> 19 October 2018, 07:34:49 UTC
1e14d3c Fix reqest count (#214) * fix manifest examples Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix MetricsCollector instance Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * eval req count after status check Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix cont check when ReqestCount is not set Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 17 October 2018, 06:30:02 UTC
44bd27e enlarge max of check goal grpc (#200) Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 17 October 2018, 06:04:00 UTC
eb12212 fix manifest examples (#213) * fix manifest examples Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> * fix MetricsCollector instance Signed-off-by: YujiOshima <yuji.oshima0x3fd@gmail.com> 17 October 2018, 05:35:43 UTC
7ad8cfc Use camel-case instead of snake-case (#204) * Use camel-case instead of snake-case * Capitalize abbreviations in variables 15 October 2018, 17:02:23 UTC
back to top