Azure
/
azureml-examples
зеркало из https://github.com/Azure/azureml-examples.git
1
0
Форкнуть 0
Код Задачи Пакеты Проекты Релизы Вики Активность
azureml-examples/README.md

39 строки
1.4 KiB
Markdown
Исходник Обычный вид История

Minimal YAML frontmatter (#387) * Minimal YAML frontmatter * Modification via prefix.md, not direct
2021-03-09 02:51:22 +03:00
---
page_type: sample
languages:
Sample browser metadata explore (#388) * Added description key to frontmatter * Wrap frontmatter yaml in XML comment, trying to avoid ugly table at top of page.
2021-03-11 05:07:07 +03:00
- azurecli
refactor for CLI 2.0 preview (#495) * scaffolding * fixes, might break * update contributing * fix * try 0.63 * fix readme * order of cells * fix env * uncomment lines * try a job * add job test * try miniconda image * try mcr image * -n -> --name for az ml workspace create * fix workspace create * same * isolate rg * fix filename * simply env; increment version * add mlflow * fix enviornment, increment version * switch to .sh * add workspace create * quick fix * fix typo * add asset creation * debugging * fix action * fix * typo * fix typo * workflow edits * fix? * edits * fix data? * Add more commands * try something * remove notebook * update * fix? * try data workaround * fix training script * add azureml-mlflow to env * saving * try sweep job * sweep edits * fix yaml * remove limits * testing * edits * minor edits to basic.yml * add exp name * saving * save; green? * move --web to untested.sh * hmm * add model assets * remove dsstore * saving w/ model create * add endpoints dir * try something for docs * try for docs * try anon data and assets * edit * path edit * fix asset path * move stuff around * try ./ * fix mistake * save * add mount * fix * change model to use run * remove model register * data hack * test * test registered file dataset * saving * testing * whoops * change strategy * fix name * switch rg back * switch creds * switch lightgbm back * fix ws * add file: * remove deepspeed * try tensorflow * remove tensorflow * edits for docs * add verify * unisntall -> remove * -0 * saving * add check_extensions * fix mistake * saving; might break * update manage-assets and test it * whoops * minor edits * local hello world * edits * simplify workflow * try compute job * ? * try compute type * fix hello world * fix? * major refactor * remove line * comments * save * save * order of operations * saving - run sweep job * magic * switch to data - possible typos * fix typos? * ... * try dedenting for sweep * edits * names * sweep hack for now * broke stuff? * test distributed * update query logic * add exp name, :1 to env * wait for job to complete * fix user error * env to version 11 * readme edit * try something * remove redundant stuff * version * try downloading cifar * try job status check * debug bash * fixes * try stuff * add back log upload * fix bash? * copy examples/endpoints from v2-preview * run formatting * add batch endpoint (#427) * hack * move compute create to setup.sh; --name to -n for workspace create * NC6 -> NC12 * try randint sweep * remove randint from sweep; download=False for pytorch-cifar * comment out logs upload from worklow * slim iris lightgbm code * try something for cifar * fix * preprocess -> process * mlflow.pytorch log model for cifar * no mlflow, fix cifar * run formatting * relocate hello world * whoops * cifar epochs 20 -> 2 * try new env for pytorch gpu * switch pytorch to mlflow logging * cleanup cifar code * fix * write to ./model? * edit batch endpoint asset path * force add model.pkl * copy workspace assets * add some schemas, more needed * add more schemas * fix schemas * Updating the schema link to CDN endpoint (#428) * Update batch endpoint test (#429) * update schemas * fix more schemas * simplify * Revert "simplify" This reverts commit 38ec77e0ca9b54051d36699573dea04c0abb583b. * Update batch endpoint sample (#432) * test python/cli refactor (#431) * saving * Update environment.yml (#430) * ? * testing * testing * testing * cli fix * fix * saving * saving * whoops * fix readme * fix readme * whoops * minor readme edit * last readme edits * add setup workspace step * fix cli tests (#433) * test * testing * fixes * save * test * fix how-to-manage-assets * minor edit * fix * fix setup-workspace * fix readme (#434) * edits for docs (#437) * fix readme * remove sweep hack * readme fix * command => sweep in schema * saving * add back setup-workspace.sh * ? * sweep hack * Saving * edit readme asset logic * -A * minor edits for docs (#438) * fix readme * remove sweep hack * readme fix * command => sweep in schema * saving * add back setup-workspace.sh * ? * sweep hack * Saving * edit readme asset logic * -A * >- => | * missed hello world * remove old env * blind copy * yolo * edits * edits * docs edits (#440) * docs edits * save * remove upgrade * fix * reorder * save (#442) * edit training jobs (#443) * saving * minor edit * fix world language? (#444) * change tf horovod env (#445) * change tf horovod env * - * move exp name * standardize cli jobs (#446) * test * standardize training jobs * online endpoints (#435) * online endpoints * run readme.py; manual edit workflows * fix delete command * fix * fix endpoint * rerun readme Co-authored-by: Cody <54814569+lostmygithubaccount@users.noreply.github.com> * quick fix (#448) * run readme (#449) * docs updates (#450) * minor edit for docs (#451) * docs updates * quick fix * undo sweep hack (#452) * edits for docs (#453) * add hello world back * updates * minor edits * Update python-ml-basic-cpu.yml * Add REST examples (#439) * Add scripts without blob cli * Everything working for endpoint except discrimnator issue with online deployment * Working simple job flow * Use default container for data * Rename scripts, make wait_for_completion a function, fix online deployment * Add scoring to endpoint * Change structure of folders * Remove todo * Move rest to cli folder * Switch to lightgbm from xgboost * Fix wait_for_completion * Get AZURE_STORAGE_KEY from az cli * Change MIR region to westus * Remove exports from variables * Fix scoring the endpoint Co-authored-by: Trent McCormick <trmccorm@microsoft.com> * Fix closing tag of sample to match opening tag. (#455) * updates for docs (#456) * switch cli version back (#457) * / * tests for rest how-to (#454) * run readme * save * test rest with azureml-examples * fixes * run readme * Use old way of getting certain variables Co-authored-by: tmccrmck <tjmcc01@gmail.com> * Update how-to-batch-score.sh (#436) Fix batch asset path * force add files (#458) * endpoint script fix (#459) * fix endpoints * fix endpoints * fix endpoint script * fix endpoints * edits for docs (#461) * use * exp name edits * bash edits * save * save * edit * Get storage info from default datastore (#460) * simplify hello world * move RG (#463) * switch credentials (#464) * debug (#465) * debug * fix typo * fjeiwpaol tefjpao * Tweak to canary-flow shellscript (#466) * Fix `test_green` snippet `test_green` snippet was dupe of `green_10pct_traffic`. Switched to syntax described in https://azure.github.io/azureml-v2-preview/_build/html/overview/concepts/endpoints/online-endpoint/scenarios/declarative-flow.html * Results of `python readme.py` This seems erroneous to me, but... * Changed endpoint name to my-new-endpoint per rest of script * fixed typo in var name Co-authored-by: Cody <54814569+lostmygithubaccount@users.noreply.github.com> * Add managed identity shell script for online endpoint (#467) * Online endpoint managed identities tests shell script * Updated online endpoint managed identity shell scripts Co-authored-by: pansav <pansav@microsoft.com> * imperative flow (#469) * .yaml -> .yml (#470) * .yaml -> yml * update python files * edit * Add get deployment logs to rest docs (#473) * Add getLogs API call * Fix typo * minor edit to hello world (#474) * add -example to exp name * edits for docs (#476) * hacks * order of operations * edits * increment model * edits to batch * fix readme * reduce training times to 1 epoch * Update setup.sh * >- -> > (#477) * update CLI version (#480) * update CLI version * try 0.71 * use 1.0.0a2 * Gopalv/tfserving sample (#479) * WIP * adding logic to extract ACR * ran readme.py * local testing works * sed works * parametrized, added local and remote test * adding workflow file * acr login failing * adding sleep after docker run * workflows * Fixed deployment yaml for online endpoint managed identity example (#482) Co-authored-by: pansav <pansav@microsoft.com> * updates to cli-preview (#484) * testing * update dockerfiles * swithc back to conda yml * test * ? * ? * dead snakes * fix job * remove non-python jobs * test * job descriptions * asset descritptions * docs * save * saveing * Fixed online endpoint automation failures issues (#489) * Fixed automation failures issues * Removed try catch block from run method * Fixed declarative yaml * Updated docker image uri to Ubuntu 18.04 * Update system identity flow to use endpoint show to retrieve object id Co-authored-by: pansav <pansav@microsoft.com> * Add how-to-train-autorest with Python SDK (#481) * Add autorest example * Fixes for REST scripts * Removed linux specific code in shell script (#492) * Removed linux specific code in shell script * Added delay after system identity is created and also changed secrets id * Changed CLI version for managed identity tests * Fixed declarative yaml file and upgraded to latest CLI version for managed identity * Changed user identity place holder name * Changed online endpoint schema url in yaml files * Fixed endpoint delete command in declarative flow * Fixed file path in declarative sample Co-authored-by: pansav <pansav@microsoft.com> * Fix how-to-deploy-script (#491) * Fix how-to-deploy-script * Use random versions for environmetns * Don't use headers for async operations * Use jq -r everywhere * Print async operation error * Try changing endpoint * Add mlflow model file (#486) * update batch sample (#493) * Gopalv/triton sample (#490) * WIP * adding logic to extract ACR * ran readme.py * local testing works * sed works * parametrized, added local and remote test * adding workflow file * acr login failing * adding sleep after docker run * workflows * updating yaml and script to exit properly * adding deletion * fixing typos * making base endpoint more generic * try after deletion * renaming tfserving * moved file to correct place * works until MFE error * not tested locally * new CLI drop seems to work * adding triton endpoint * using anonymous environments * adding back placeholder * tfserving fails for some reason * triton works * get different acr * install pyindex * deleting model * fixing model delete * factoring cleanup utility * adding exception handling for acr fail * fixing syntax error * making image url an argument * using public image for tfserving * add back tutorial test, rename * manual modify of recently added workflow * repin black; minor edit * run black-nb * fix typo * test horovod env (#496) * test horovod env * revert to old environment * try v2 * revert * try oss horovod * try tf image * try horoovd image * switch horovod env * 1.0.0a2 -> 1.0.0a4 * update batch sample (#497) * Removed min/max instances and changed auth_mode to key (#498) Co-authored-by: pansav <pansav@microsoft.com> * saving (#499) * minor reqs edits Co-authored-by: Tracy Chen <47156502+tracychms@users.noreply.github.com> Co-authored-by: Harneet Virk <harnvir@microsoft.com> Co-authored-by: Sethu Raman <rsethur@users.noreply.github.com> Co-authored-by: trent <tjmcc01@gmail.com> Co-authored-by: Trent McCormick <trmccorm@microsoft.com> Co-authored-by: vmagelo <51368918+vmagelo@users.noreply.github.com> Co-authored-by: Larry O'Brien <lobrien@knowing.net> Co-authored-by: psavdekar <45885005+psavdekar@users.noreply.github.com> Co-authored-by: pansav <pansav@microsoft.com> Co-authored-by: Gopal Vashishtha <gopalv@microsoft.com>
2021-05-12 04:53:32 +03:00
- python
Minimal YAML frontmatter (#387) * Minimal YAML frontmatter * Modification via prefix.md, not direct
2021-03-09 02:51:22 +03:00
products:
- azure-machine-learning
refactor for CLI 2.0 preview (#495) * scaffolding * fixes, might break * update contributing * fix * try 0.63 * fix readme * order of cells * fix env * uncomment lines * try a job * add job test * try miniconda image * try mcr image * -n -> --name for az ml workspace create * fix workspace create * same * isolate rg * fix filename * simply env; increment version * add mlflow * fix enviornment, increment version * switch to .sh * add workspace create * quick fix * fix typo * add asset creation * debugging * fix action * fix * typo * fix typo * workflow edits * fix? * edits * fix data? * Add more commands * try something * remove notebook * update * fix? * try data workaround * fix training script * add azureml-mlflow to env * saving * try sweep job * sweep edits * fix yaml * remove limits * testing * edits * minor edits to basic.yml * add exp name * saving * save; green? * move --web to untested.sh * hmm * add model assets * remove dsstore * saving w/ model create * add endpoints dir * try something for docs * try for docs * try anon data and assets * edit * path edit * fix asset path * move stuff around * try ./ * fix mistake * save * add mount * fix * change model to use run * remove model register * data hack * test * test registered file dataset * saving * testing * whoops * change strategy * fix name * switch rg back * switch creds * switch lightgbm back * fix ws * add file: * remove deepspeed * try tensorflow * remove tensorflow * edits for docs * add verify * unisntall -> remove * -0 * saving * add check_extensions * fix mistake * saving; might break * update manage-assets and test it * whoops * minor edits * local hello world * edits * simplify workflow * try compute job * ? * try compute type * fix hello world * fix? * major refactor * remove line * comments * save * save * order of operations * saving - run sweep job * magic * switch to data - possible typos * fix typos? * ... * try dedenting for sweep * edits * names * sweep hack for now * broke stuff? * test distributed * update query logic * add exp name, :1 to env * wait for job to complete * fix user error * env to version 11 * readme edit * try something * remove redundant stuff * version * try downloading cifar * try job status check * debug bash * fixes * try stuff * add back log upload * fix bash? * copy examples/endpoints from v2-preview * run formatting * add batch endpoint (#427) * hack * move compute create to setup.sh; --name to -n for workspace create * NC6 -> NC12 * try randint sweep * remove randint from sweep; download=False for pytorch-cifar * comment out logs upload from worklow * slim iris lightgbm code * try something for cifar * fix * preprocess -> process * mlflow.pytorch log model for cifar * no mlflow, fix cifar * run formatting * relocate hello world * whoops * cifar epochs 20 -> 2 * try new env for pytorch gpu * switch pytorch to mlflow logging * cleanup cifar code * fix * write to ./model? * edit batch endpoint asset path * force add model.pkl * copy workspace assets * add some schemas, more needed * add more schemas * fix schemas * Updating the schema link to CDN endpoint (#428) * Update batch endpoint test (#429) * update schemas * fix more schemas * simplify * Revert "simplify" This reverts commit 38ec77e0ca9b54051d36699573dea04c0abb583b. * Update batch endpoint sample (#432) * test python/cli refactor (#431) * saving * Update environment.yml (#430) * ? * testing * testing * testing * cli fix * fix * saving * saving * whoops * fix readme * fix readme * whoops * minor readme edit * last readme edits * add setup workspace step * fix cli tests (#433) * test * testing * fixes * save * test * fix how-to-manage-assets * minor edit * fix * fix setup-workspace * fix readme (#434) * edits for docs (#437) * fix readme * remove sweep hack * readme fix * command => sweep in schema * saving * add back setup-workspace.sh * ? * sweep hack * Saving * edit readme asset logic * -A * minor edits for docs (#438) * fix readme * remove sweep hack * readme fix * command => sweep in schema * saving * add back setup-workspace.sh * ? * sweep hack * Saving * edit readme asset logic * -A * >- => | * missed hello world * remove old env * blind copy * yolo * edits * edits * docs edits (#440) * docs edits * save * remove upgrade * fix * reorder * save (#442) * edit training jobs (#443) * saving * minor edit * fix world language? (#444) * change tf horovod env (#445) * change tf horovod env * - * move exp name * standardize cli jobs (#446) * test * standardize training jobs * online endpoints (#435) * online endpoints * run readme.py; manual edit workflows * fix delete command * fix * fix endpoint * rerun readme Co-authored-by: Cody <54814569+lostmygithubaccount@users.noreply.github.com> * quick fix (#448) * run readme (#449) * docs updates (#450) * minor edit for docs (#451) * docs updates * quick fix * undo sweep hack (#452) * edits for docs (#453) * add hello world back * updates * minor edits * Update python-ml-basic-cpu.yml * Add REST examples (#439) * Add scripts without blob cli * Everything working for endpoint except discrimnator issue with online deployment * Working simple job flow * Use default container for data * Rename scripts, make wait_for_completion a function, fix online deployment * Add scoring to endpoint * Change structure of folders * Remove todo * Move rest to cli folder * Switch to lightgbm from xgboost * Fix wait_for_completion * Get AZURE_STORAGE_KEY from az cli * Change MIR region to westus * Remove exports from variables * Fix scoring the endpoint Co-authored-by: Trent McCormick <trmccorm@microsoft.com> * Fix closing tag of sample to match opening tag. (#455) * updates for docs (#456) * switch cli version back (#457) * / * tests for rest how-to (#454) * run readme * save * test rest with azureml-examples * fixes * run readme * Use old way of getting certain variables Co-authored-by: tmccrmck <tjmcc01@gmail.com> * Update how-to-batch-score.sh (#436) Fix batch asset path * force add files (#458) * endpoint script fix (#459) * fix endpoints * fix endpoints * fix endpoint script * fix endpoints * edits for docs (#461) * use * exp name edits * bash edits * save * save * edit * Get storage info from default datastore (#460) * simplify hello world * move RG (#463) * switch credentials (#464) * debug (#465) * debug * fix typo * fjeiwpaol tefjpao * Tweak to canary-flow shellscript (#466) * Fix `test_green` snippet `test_green` snippet was dupe of `green_10pct_traffic`. Switched to syntax described in https://azure.github.io/azureml-v2-preview/_build/html/overview/concepts/endpoints/online-endpoint/scenarios/declarative-flow.html * Results of `python readme.py` This seems erroneous to me, but... * Changed endpoint name to my-new-endpoint per rest of script * fixed typo in var name Co-authored-by: Cody <54814569+lostmygithubaccount@users.noreply.github.com> * Add managed identity shell script for online endpoint (#467) * Online endpoint managed identities tests shell script * Updated online endpoint managed identity shell scripts Co-authored-by: pansav <pansav@microsoft.com> * imperative flow (#469) * .yaml -> .yml (#470) * .yaml -> yml * update python files * edit * Add get deployment logs to rest docs (#473) * Add getLogs API call * Fix typo * minor edit to hello world (#474) * add -example to exp name * edits for docs (#476) * hacks * order of operations * edits * increment model * edits to batch * fix readme * reduce training times to 1 epoch * Update setup.sh * >- -> > (#477) * update CLI version (#480) * update CLI version * try 0.71 * use 1.0.0a2 * Gopalv/tfserving sample (#479) * WIP * adding logic to extract ACR * ran readme.py * local testing works * sed works * parametrized, added local and remote test * adding workflow file * acr login failing * adding sleep after docker run * workflows * Fixed deployment yaml for online endpoint managed identity example (#482) Co-authored-by: pansav <pansav@microsoft.com> * updates to cli-preview (#484) * testing * update dockerfiles * swithc back to conda yml * test * ? * ? * dead snakes * fix job * remove non-python jobs * test * job descriptions * asset descritptions * docs * save * saveing * Fixed online endpoint automation failures issues (#489) * Fixed automation failures issues * Removed try catch block from run method * Fixed declarative yaml * Updated docker image uri to Ubuntu 18.04 * Update system identity flow to use endpoint show to retrieve object id Co-authored-by: pansav <pansav@microsoft.com> * Add how-to-train-autorest with Python SDK (#481) * Add autorest example * Fixes for REST scripts * Removed linux specific code in shell script (#492) * Removed linux specific code in shell script * Added delay after system identity is created and also changed secrets id * Changed CLI version for managed identity tests * Fixed declarative yaml file and upgraded to latest CLI version for managed identity * Changed user identity place holder name * Changed online endpoint schema url in yaml files * Fixed endpoint delete command in declarative flow * Fixed file path in declarative sample Co-authored-by: pansav <pansav@microsoft.com> * Fix how-to-deploy-script (#491) * Fix how-to-deploy-script * Use random versions for environmetns * Don't use headers for async operations * Use jq -r everywhere * Print async operation error * Try changing endpoint * Add mlflow model file (#486) * update batch sample (#493) * Gopalv/triton sample (#490) * WIP * adding logic to extract ACR * ran readme.py * local testing works * sed works * parametrized, added local and remote test * adding workflow file * acr login failing * adding sleep after docker run * workflows * updating yaml and script to exit properly * adding deletion * fixing typos * making base endpoint more generic * try after deletion * renaming tfserving * moved file to correct place * works until MFE error * not tested locally * new CLI drop seems to work * adding triton endpoint * using anonymous environments * adding back placeholder * tfserving fails for some reason * triton works * get different acr * install pyindex * deleting model * fixing model delete * factoring cleanup utility * adding exception handling for acr fail * fixing syntax error * making image url an argument * using public image for tfserving * add back tutorial test, rename * manual modify of recently added workflow * repin black; minor edit * run black-nb * fix typo * test horovod env (#496) * test horovod env * revert to old environment * try v2 * revert * try oss horovod * try tf image * try horoovd image * switch horovod env * 1.0.0a2 -> 1.0.0a4 * update batch sample (#497) * Removed min/max instances and changed auth_mode to key (#498) Co-authored-by: pansav <pansav@microsoft.com> * saving (#499) * minor reqs edits Co-authored-by: Tracy Chen <47156502+tracychms@users.noreply.github.com> Co-authored-by: Harneet Virk <harnvir@microsoft.com> Co-authored-by: Sethu Raman <rsethur@users.noreply.github.com> Co-authored-by: trent <tjmcc01@gmail.com> Co-authored-by: Trent McCormick <trmccorm@microsoft.com> Co-authored-by: vmagelo <51368918+vmagelo@users.noreply.github.com> Co-authored-by: Larry O'Brien <lobrien@knowing.net> Co-authored-by: psavdekar <45885005+psavdekar@users.noreply.github.com> Co-authored-by: pansav <pansav@microsoft.com> Co-authored-by: Gopal Vashishtha <gopalv@microsoft.com>
2021-05-12 04:53:32 +03:00
description: Top-level directory for official Azure Machine Learning sample code.
Minimal YAML frontmatter (#387) * Minimal YAML frontmatter * Modification via prefix.md, not direct
2021-03-09 02:51:22 +03:00
---
Sample browser metadata explore (#388) * Added description key to frontmatter * Wrap frontmatter yaml in XML comment, trying to avoid ugly table at top of page.
2021-03-11 05:07:07 +03:00
minor readme edits (#401) * readme edits * remove lightgbm tutorial
2021-03-24 23:29:42 +03:00
# Azure Machine Learning Examples
reset git history - final
2020-11-11 23:56:18 +03:00
add smoke badge to readme (#351)
2021-02-01 06:25:10 +03:00
[![smoke](https://github.com/Azure/azureml-examples/workflows/smoke/badge.svg)](https://github.com/Azure/azureml-examples/actions?query=workflow%3Asmoke)
reset git history - final
2020-11-11 23:56:18 +03:00
[![cleanup](https://github.com/Azure/azureml-examples/workflows/cleanup/badge.svg)](https://github.com/Azure/azureml-examples/actions?query=workflow%3Acleanup)
[![code style: black](https://img.shields.io/badge/code%20style-black-000000.svg)](https://github.com/psf/black)
[![license: MIT](https://img.shields.io/badge/License-MIT-purple.svg)](LICENSE)
refactor readme.py and workflow logic (#342) * saving * prefix edit * save for later * save for later * fix typos * logic * fix and rerun readme * add dataprep to reqs * remove papermill -k * add azureml-train back to reqs * try this * revert and pin papermill version * edit workflows * same to other workflows * revert * blaming vim * out of ideas * try this * facepalm? * run tutorials too * edit prefix
2021-02-01 04:13:21 +03:00
Welcome to the Azure Machine Learning examples repository!
reset git history - final
2020-11-11 23:56:18 +03:00
readme updates; minor requirements.txt updates (#390) * saving; need to run readme * update requirements.txt; run readme.py * move contents to suffix * add blurb * minor edit
2021-03-12 23:49:12 +03:00
## Contents
|directory|description|
|-|-|
|`.github`|GitHub specific files like Actions workflow yaml definitions and issue templates|
refactor for CLI 2.0 preview (#495) * scaffolding * fixes, might break * update contributing * fix * try 0.63 * fix readme * order of cells * fix env * uncomment lines * try a job * add job test * try miniconda image * try mcr image * -n -> --name for az ml workspace create * fix workspace create * same * isolate rg * fix filename * simply env; increment version * add mlflow * fix enviornment, increment version * switch to .sh * add workspace create * quick fix * fix typo * add asset creation * debugging * fix action * fix * typo * fix typo * workflow edits * fix? * edits * fix data? * Add more commands * try something * remove notebook * update * fix? * try data workaround * fix training script * add azureml-mlflow to env * saving * try sweep job * sweep edits * fix yaml * remove limits * testing * edits * minor edits to basic.yml * add exp name * saving * save; green? * move --web to untested.sh * hmm * add model assets * remove dsstore * saving w/ model create * add endpoints dir * try something for docs * try for docs * try anon data and assets * edit * path edit * fix asset path * move stuff around * try ./ * fix mistake * save * add mount * fix * change model to use run * remove model register * data hack * test * test registered file dataset * saving * testing * whoops * change strategy * fix name * switch rg back * switch creds * switch lightgbm back * fix ws * add file: * remove deepspeed * try tensorflow * remove tensorflow * edits for docs * add verify * unisntall -> remove * -0 * saving * add check_extensions * fix mistake * saving; might break * update manage-assets and test it * whoops * minor edits * local hello world * edits * simplify workflow * try compute job * ? * try compute type * fix hello world * fix? * major refactor * remove line * comments * save * save * order of operations * saving - run sweep job * magic * switch to data - possible typos * fix typos? * ... * try dedenting for sweep * edits * names * sweep hack for now * broke stuff? * test distributed * update query logic * add exp name, :1 to env * wait for job to complete * fix user error * env to version 11 * readme edit * try something * remove redundant stuff * version * try downloading cifar * try job status check * debug bash * fixes * try stuff * add back log upload * fix bash? * copy examples/endpoints from v2-preview * run formatting * add batch endpoint (#427) * hack * move compute create to setup.sh; --name to -n for workspace create * NC6 -> NC12 * try randint sweep * remove randint from sweep; download=False for pytorch-cifar * comment out logs upload from worklow * slim iris lightgbm code * try something for cifar * fix * preprocess -> process * mlflow.pytorch log model for cifar * no mlflow, fix cifar * run formatting * relocate hello world * whoops * cifar epochs 20 -> 2 * try new env for pytorch gpu * switch pytorch to mlflow logging * cleanup cifar code * fix * write to ./model? * edit batch endpoint asset path * force add model.pkl * copy workspace assets * add some schemas, more needed * add more schemas * fix schemas * Updating the schema link to CDN endpoint (#428) * Update batch endpoint test (#429) * update schemas * fix more schemas * simplify * Revert "simplify" This reverts commit 38ec77e0ca9b54051d36699573dea04c0abb583b. * Update batch endpoint sample (#432) * test python/cli refactor (#431) * saving * Update environment.yml (#430) * ? * testing * testing * testing * cli fix * fix * saving * saving * whoops * fix readme * fix readme * whoops * minor readme edit * last readme edits * add setup workspace step * fix cli tests (#433) * test * testing * fixes * save * test * fix how-to-manage-assets * minor edit * fix * fix setup-workspace * fix readme (#434) * edits for docs (#437) * fix readme * remove sweep hack * readme fix * command => sweep in schema * saving * add back setup-workspace.sh * ? * sweep hack * Saving * edit readme asset logic * -A * minor edits for docs (#438) * fix readme * remove sweep hack * readme fix * command => sweep in schema * saving * add back setup-workspace.sh * ? * sweep hack * Saving * edit readme asset logic * -A * >- => | * missed hello world * remove old env * blind copy * yolo * edits * edits * docs edits (#440) * docs edits * save * remove upgrade * fix * reorder * save (#442) * edit training jobs (#443) * saving * minor edit * fix world language? (#444) * change tf horovod env (#445) * change tf horovod env * - * move exp name * standardize cli jobs (#446) * test * standardize training jobs * online endpoints (#435) * online endpoints * run readme.py; manual edit workflows * fix delete command * fix * fix endpoint * rerun readme Co-authored-by: Cody <54814569+lostmygithubaccount@users.noreply.github.com> * quick fix (#448) * run readme (#449) * docs updates (#450) * minor edit for docs (#451) * docs updates * quick fix * undo sweep hack (#452) * edits for docs (#453) * add hello world back * updates * minor edits * Update python-ml-basic-cpu.yml * Add REST examples (#439) * Add scripts without blob cli * Everything working for endpoint except discrimnator issue with online deployment * Working simple job flow * Use default container for data * Rename scripts, make wait_for_completion a function, fix online deployment * Add scoring to endpoint * Change structure of folders * Remove todo * Move rest to cli folder * Switch to lightgbm from xgboost * Fix wait_for_completion * Get AZURE_STORAGE_KEY from az cli * Change MIR region to westus * Remove exports from variables * Fix scoring the endpoint Co-authored-by: Trent McCormick <trmccorm@microsoft.com> * Fix closing tag of sample to match opening tag. (#455) * updates for docs (#456) * switch cli version back (#457) * / * tests for rest how-to (#454) * run readme * save * test rest with azureml-examples * fixes * run readme * Use old way of getting certain variables Co-authored-by: tmccrmck <tjmcc01@gmail.com> * Update how-to-batch-score.sh (#436) Fix batch asset path * force add files (#458) * endpoint script fix (#459) * fix endpoints * fix endpoints * fix endpoint script * fix endpoints * edits for docs (#461) * use * exp name edits * bash edits * save * save * edit * Get storage info from default datastore (#460) * simplify hello world * move RG (#463) * switch credentials (#464) * debug (#465) * debug * fix typo * fjeiwpaol tefjpao * Tweak to canary-flow shellscript (#466) * Fix `test_green` snippet `test_green` snippet was dupe of `green_10pct_traffic`. Switched to syntax described in https://azure.github.io/azureml-v2-preview/_build/html/overview/concepts/endpoints/online-endpoint/scenarios/declarative-flow.html * Results of `python readme.py` This seems erroneous to me, but... * Changed endpoint name to my-new-endpoint per rest of script * fixed typo in var name Co-authored-by: Cody <54814569+lostmygithubaccount@users.noreply.github.com> * Add managed identity shell script for online endpoint (#467) * Online endpoint managed identities tests shell script * Updated online endpoint managed identity shell scripts Co-authored-by: pansav <pansav@microsoft.com> * imperative flow (#469) * .yaml -> .yml (#470) * .yaml -> yml * update python files * edit * Add get deployment logs to rest docs (#473) * Add getLogs API call * Fix typo * minor edit to hello world (#474) * add -example to exp name * edits for docs (#476) * hacks * order of operations * edits * increment model * edits to batch * fix readme * reduce training times to 1 epoch * Update setup.sh * >- -> > (#477) * update CLI version (#480) * update CLI version * try 0.71 * use 1.0.0a2 * Gopalv/tfserving sample (#479) * WIP * adding logic to extract ACR * ran readme.py * local testing works * sed works * parametrized, added local and remote test * adding workflow file * acr login failing * adding sleep after docker run * workflows * Fixed deployment yaml for online endpoint managed identity example (#482) Co-authored-by: pansav <pansav@microsoft.com> * updates to cli-preview (#484) * testing * update dockerfiles * swithc back to conda yml * test * ? * ? * dead snakes * fix job * remove non-python jobs * test * job descriptions * asset descritptions * docs * save * saveing * Fixed online endpoint automation failures issues (#489) * Fixed automation failures issues * Removed try catch block from run method * Fixed declarative yaml * Updated docker image uri to Ubuntu 18.04 * Update system identity flow to use endpoint show to retrieve object id Co-authored-by: pansav <pansav@microsoft.com> * Add how-to-train-autorest with Python SDK (#481) * Add autorest example * Fixes for REST scripts * Removed linux specific code in shell script (#492) * Removed linux specific code in shell script * Added delay after system identity is created and also changed secrets id * Changed CLI version for managed identity tests * Fixed declarative yaml file and upgraded to latest CLI version for managed identity * Changed user identity place holder name * Changed online endpoint schema url in yaml files * Fixed endpoint delete command in declarative flow * Fixed file path in declarative sample Co-authored-by: pansav <pansav@microsoft.com> * Fix how-to-deploy-script (#491) * Fix how-to-deploy-script * Use random versions for environmetns * Don't use headers for async operations * Use jq -r everywhere * Print async operation error * Try changing endpoint * Add mlflow model file (#486) * update batch sample (#493) * Gopalv/triton sample (#490) * WIP * adding logic to extract ACR * ran readme.py * local testing works * sed works * parametrized, added local and remote test * adding workflow file * acr login failing * adding sleep after docker run * workflows * updating yaml and script to exit properly * adding deletion * fixing typos * making base endpoint more generic * try after deletion * renaming tfserving * moved file to correct place * works until MFE error * not tested locally * new CLI drop seems to work * adding triton endpoint * using anonymous environments * adding back placeholder * tfserving fails for some reason * triton works * get different acr * install pyindex * deleting model * fixing model delete * factoring cleanup utility * adding exception handling for acr fail * fixing syntax error * making image url an argument * using public image for tfserving * add back tutorial test, rename * manual modify of recently added workflow * repin black; minor edit * run black-nb * fix typo * test horovod env (#496) * test horovod env * revert to old environment * try v2 * revert * try oss horovod * try tf image * try horoovd image * switch horovod env * 1.0.0a2 -> 1.0.0a4 * update batch sample (#497) * Removed min/max instances and changed auth_mode to key (#498) Co-authored-by: pansav <pansav@microsoft.com> * saving (#499) * minor reqs edits Co-authored-by: Tracy Chen <47156502+tracychms@users.noreply.github.com> Co-authored-by: Harneet Virk <harnvir@microsoft.com> Co-authored-by: Sethu Raman <rsethur@users.noreply.github.com> Co-authored-by: trent <tjmcc01@gmail.com> Co-authored-by: Trent McCormick <trmccorm@microsoft.com> Co-authored-by: vmagelo <51368918+vmagelo@users.noreply.github.com> Co-authored-by: Larry O'Brien <lobrien@knowing.net> Co-authored-by: psavdekar <45885005+psavdekar@users.noreply.github.com> Co-authored-by: pansav <pansav@microsoft.com> Co-authored-by: Gopal Vashishtha <gopalv@microsoft.com>
2021-05-12 04:53:32 +03:00
|`cli`|Azure Machine Learning CLI 2.0 (preview) examples|
|`python-sdk`|Azure Machine Learning Python SDK examples|
readme updates; minor requirements.txt updates (#390) * saving; need to run readme * update requirements.txt; run readme.py * move contents to suffix * add blurb * minor edit
2021-03-12 23:49:12 +03:00
minor readme edits (#289) * more minor edits to readme * revert
2020-11-16 23:52:16 +03:00
## Contributing
We welcome contributions and suggestions! Please see the [contributing guidelines](CONTRIBUTING.md) for details.
## Code of Conduct
This project has adopted the [Microsoft Open Source Code of Conduct](https://opensource.microsoft.com/codeofconduct/). Please see the [code of conduct](CODE_OF_CONDUCT.md) for details.
reset git history - final
2020-11-11 23:56:18 +03:00
## Reference
minor readme edits (#401) * readme edits * remove lightgbm tutorial
2021-03-24 23:29:42 +03:00
- [Documentation](https://docs.microsoft.com/azure/machine-learning)