feat: hub search and metadata, trainer job handles, typed deploy options, endpoint diagnostics, Bedrock and S3 helpers by evakravi · Pull Request #6533 · aws/sagemaker-python-sdk

feat: expose the created job as latest_training_job and return it from ModelTrainer.train()

evakravi

committed

450c866

fix: make ModelBuilder.deploy(tags=...) a named parameter that reaches every deploy path

evakravi

committed

6e9df41

feat(core): add HubContent.get_all and keep OriginalCreationTime on listed hub content

evakravi

committed

c2db676

feat(core): let HubContent read its own metadata from search keywords and document

evakravi

committed

d352fe5

feat(train): call on_submit with the created TrainingJob before the wait

evakravi

committed

8c8dd9a

feat(serve): name the ModelBuilder.deploy endpoint options and reject an unknown kwarg

evakravi

committed

b04f4c9

docs: fix the managed spot example in the migration guide

evakravi

committed

b6959d8

feat(core): add HubContent.search with Metadata filters on the Search API

evakravi

committed

93c0a67

fix(core): accept (key, value) pairs in format_tags again

evakravi

committed

097192a

feat(train): add get_metric_history for MLflow metric data

evakravi

committed

fb7c616

feat(train): type the reward-function contract and use one Lambda runtime

evakravi

committed

a4c2786

feat(train): validate LLM-as-Judge custom metrics before the job starts

evakravi

committed

5a134b9

feat(core): add S3Downloader byte-range reads, line streaming, and head metadata

evakravi

committed

2c09cf5

feat(train): detect a dataset format from a local or S3 sample

evakravi

committed

cd59491

feat(serve): add Bedrock model listing and an invocable deployment handle

evakravi

committed

41aae12

feat(core): describe SageMaker instance types through one public helper

evakravi

committed

d906911

feat(core): add ModelPackageGroup.get_or_create

evakravi

committed

6e820f6

fix(core): return a session-bound ProcessingJob with its ARN from Processor._start_new

evakravi

committed

0727c47

feat(core): add raise_on_stopped to every generated wait() method

evakravi

committed

c6294e1

feat(core): add Endpoint.diagnostics and Endpoint.recent_logs

evakravi

committed

4c9bedb

feat(train): add FineTuningSweep for serverless trainer parameter sweeps

evakravi

committed

4b89520

fix: address review comments on get_all docstrings, format_tags typing, and the tag merge

evakravi

committed

a20a1e6

Navigation Menu

Repository navigation

feat: hub search and metadata, trainer job handles, typed deploy options, endpoint diagnostics, Bedrock and S3 helpers - #6533

feat: hub search and metadata, trainer job handles, typed deploy options, endpoint diagnostics, Bedrock and S3 helpers#6533
evakravi wants to merge 30 commits into
aws:masterfrom
evakravi:feat/hub-content-trainer-builder

Commits on Oct 23, 2026

feat: expose the created job as latest_training_job and return it from ModelTrainer.train()

fix: make ModelBuilder.deploy(tags=...) a named parameter that reaches every deploy path

feat(core): add HubContent.get_all and keep OriginalCreationTime on listed hub content

feat(core): let HubContent read its own metadata from search keywords and document

feat(train): call on_submit with the created TrainingJob before the wait

feat(serve): name the ModelBuilder.deploy endpoint options and reject an unknown kwarg

docs: fix the managed spot example in the migration guide

feat(core): add HubContent.search with Metadata filters on the Search API

fix(core): accept (key, value) pairs in format_tags again

feat(train): add get_metric_history for MLflow metric data

feat(train): type the reward-function contract and use one Lambda runtime

feat(train): validate LLM-as-Judge custom metrics before the job starts

feat(core): add S3Downloader byte-range reads, line streaming, and head metadata

feat(train): detect a dataset format from a local or S3 sample

feat(serve): add Bedrock model listing and an invocable deployment handle

feat(core): describe SageMaker instance types through one public helper

feat(core): add ModelPackageGroup.get_or_create

fix(core): return a session-bound ProcessingJob with its ARN from Processor._start_new

feat(core): add raise_on_stopped to every generated wait() method

feat(core): add Endpoint.diagnostics and Endpoint.recent_logs

feat(train): add FineTuningSweep for serverless trainer parameter sweeps

fix: address review comments on get_all docstrings, format_tags typing, and the tag merge

Commits on Oct 24, 2026

chore(merge): merge origin/master into feat/hub-content-trainer-builder

fix(codegen): make the generated modules pass black-check and pydocstyle

fix: scope deploy() tags to the call and stringify every tag representation

Commits on Oct 26, 2026

feat(core): describe listed hub contents concurrently through ResourceIterator max_workers

feat(core): add HubContent.get_all_cached with a local cache of described hub contents

feat(core): emit telemetry from HubContent.search, get_all_cached, and refresh_cache

refactor(core): build the hub catalog cache on LRUCache, the JumpStart cache shape

feat(core): serve HubContent.get_all from the hub catalog cache on the default thread pool