Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Chronos: train tcn model on gpu and speed up inference on cpu #5594

Merged
merged 6 commits into from
Sep 9, 2022
Merged

Chronos: train tcn model on gpu and speed up inference on cpu #5594

merged 6 commits into from
Sep 9, 2022

Conversation

smurf-1119
Copy link
Contributor

@smurf-1119 smurf-1119 commented Aug 31, 2022 •
edited by TheaperDeng
Loading

Description

uncompleted
train tcn model on gpu and speed up inference on cpu

https://qp-bigdl.readthedocs.io/en/nano-gpu/doc/Chronos/QuickStart/index.html

@TheaperDeng TheaperDeng changed the title []train tcn model on gpu and speed up inference on cpu [WIP] train tcn model on gpu and speed up inference on cpu Aug 31, 2022
@TheaperDeng
Copy link
Contributor

  1. It should be better if we could use

to wrap TCN model

  1. Merge the scritps from train and predict to 1 script where

1). create a model

2). train the model on gpu

3). trace the model(Trainer.trace)

4). test the latency (Use model(x) and Evaluator.get_latencywhere x is one sample data)

Some reference link:

[Chronos-workshop/chronos_forecasting.ipynb at main · intel-analytics/Chronos-workshop · GitHub](https://github.com/intel-analytics/Chronos-workshop/blob/main/notebook/chronos_forecasting.ipynb)

https://bigdl.readthedocs.io/en/latest/doc/PythonAPI/Chronos/evaluator.html#bigdl.chronos.metric.forecast_metrics.Evaluator.get_latency

TheaperDeng
TheaperDeng reviewed Sep 2, 2022
View reviewed changes
python/chronos/example/nano_qp/tcn/train_test.py Outdated
@@ -0,0 +1,91 @@
from pytorch_lightning.callbacks import ModelCheckpoint
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ModelCheckpoint is not used

python/chronos/example/nano_qp/tcn/train_test.py Outdated
@@ -0,0 +1,91 @@
from pytorch_lightning.callbacks import ModelCheckpoint
import pytorch_lightning as pl
from model import LitTCN
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LitTCN is not used

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

pls fix this

python/chronos/example/nano_qp/tcn/train_test.py Outdated
from pytorch_lightning.callbacks import ModelCheckpoint
import pytorch_lightning as pl
from model import LitTCN
import time
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

not used

python/chronos/example/nano_qp/tcn/train_test.py Outdated
from bigdl.nano.pytorch import Trainer
from bigdl.chronos.model.tcn import model_creator
from bigdl.chronos.metric.forecast_metrics import Evaluator
from tcn_config import *
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

tcn_config is deleted

python/chronos/example/nano_qp/tcn/train_test.py Outdated
return tsdata_traindataloader,\
tsdata_valdataloader,\
tsdata_testdataloader,\
tsdata_test
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

tsdata_test is never used

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

pls fix this

python/chronos/example/nano_qp/tcn/train_test.py Outdated Show resolved Hide resolved
python/chronos/example/nano_qp/tcn/train_test.py Outdated

def test_time(model, dataloader):
for x, _ in dataloader:
res = model(x[0].unsqueeze(0))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more. 

def predict_wraper(model, input_sample):
    model(input_sample)

python/chronos/example/nano_qp/tcn/train_test.py Outdated

print("original pytorch runtime (ms):", Evaluator.get_latency(test_time,lit_model, tsdata_testdataloader))

speed_model = Trainer.trace(lit_model, accelerator="onnxruntime", input_sample=tsdata_testdataloader)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You should use tsdata_traindataloader for input_sample

python/chronos/example/nano_qp/tcn/train_test.py Outdated
print("original pytorch runtime (ms):", Evaluator.get_latency(test_time,lit_model, tsdata_testdataloader))

speed_model = Trainer.trace(lit_model, accelerator="onnxruntime", input_sample=tsdata_testdataloader)
print("onnxruntime pytorch runtime (ms):", Evaluator.get_latency(test_time, speed_model, tsdata_testdataloader))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

onnxruntime latency (ms)

python/chronos/example/nano_qp/tcn/train_test.py Outdated
trainer.fit(lit_model, tsdata_traindataloader, tsdata_testdataloader)


print("original pytorch runtime (ms):", Evaluator.get_latency(test_time,lit_model, tsdata_testdataloader))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

original pytorch latency (ms)

TheaperDeng
TheaperDeng reviewed Sep 7, 2022
View reviewed changes
python/chronos/example/nano_qp/tcn/train_test.py Outdated
print("original pytorch latency (ms):", Evaluator.get_latency(predict_wraper, lit_model, input_sample))

speed_model = Trainer.trace(lit_model, accelerator="onnxruntime", input_sample=input_sample)
print("onnxruntime latency (ms):", Evaluator.get_latency(predict_wraper, speed_model, input_sample))
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

add a new line here

python/chronos/example/nano_qp/tcn/train_test.py Outdated
@@ -0,0 +1,91 @@
from pytorch_lightning.callbacks import ModelCheckpoint
import pytorch_lightning as pl
from model import LitTCN
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

pls fix this

python/chronos/example/nano_qp/tcn/train_test.py Outdated
return tsdata_traindataloader,\
tsdata_valdataloader,\
tsdata_testdataloader,\
tsdata_test
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

pls fix this

@TheaperDeng TheaperDeng changed the title [WIP] train tcn model on gpu and speed up inference on cpu train tcn model on gpu and speed up inference on cpu Sep 7, 2022
@TheaperDeng TheaperDeng changed the title train tcn model on gpu and speed up inference on cpu Chronos: train tcn model on gpu and speed up inference on cpu Sep 9, 2022
@TheaperDeng TheaperDeng merged commit df3651b into intel-analytics:main Sep 9, 2022
ForJadeForest pushed a commit to ForJadeForest/BigDL that referenced this pull request Sep 20, 2022
@smurf-1119 @TheaperDeng @ForJadeForest
Chronos: train tcn model on gpu and speed up inference on cpu (intel-…
aaae407 
…analytics#5594)

* add nano_gpu

* simplify the code

* simplify the code

* modify the code

* add some updates

* add document

Co-authored-by: theaperdeng <[email protected]>
hkvision added a commit that referenced this pull request Sep 20, 2022
@ForJadeForest @Hanyu-Jin
@Litchilitchy @y199387 @zhentaocc @lalalapotter @AndyLuo1029 @plusbang @dding3 @sgwhat @liu-shaojun @yexinyinancy @gc-fu @Zephyr596 @rnwang04 @ShanSimu @hzjane @xiangyuT @Uxito-Ada @leonardozcm @hkvision @pinggao187 @zzti-bsj @shanyu-sys @aixideng @liangs6212 @qiyuangong @jason-dai @jenniew @PatrickkZ @Oscilloscope98 @TheaperDeng @Chaselzxy @Le-Zheng @Cyberpunk1210 @MeouSker77 @songhappy @hjzin @glorysdj @smurf-1119 @cyita @devWangBin @mycprotein
Remove rl_pong example from Jenkins example test (#5796)
492e05c 
* Enable FL Server in SGX (#5396)

* add assembly

* Enable FL Sever in SGX

* Change the url in tutorial.md

* Add start-fl-server.py to Docker Image

* rm ppml-assebly.xml

* Add start-python-fl-server-sgx.sh to Docker Image

* Put start-fl-server.py into assembly

* Update 3.1.2 Start FL Server in SGX in pytorch-nn-lr-tutorial.md

* Change the location of start-fl-server.py

* Correct the path in start-python-fl-server-sgx.sh

* Delete -r in cp command

Co-authored-by: Jiaming <[email protected]>

* fix tf inc metric format (#5127)

* fix tf inc metric format

* re-run actions

* Apply suggestions from code review

Co-authored-by: Zhentao Chen <[email protected]>

* Update

Co-authored-by: Zhentao Chen <[email protected]>

* Orca: update known-issues. (#5505)

* fix: update known-issues.

* fix: fix wording.

* Chronos: update js, css and md files to add filter function in tutorial page. (#5506)

* Chronos: PR validation job to Github action (#5349)

Migrate current jenkins CI/CD jobs to github action, this is chronos PR validation-Spark2.4-ray job

* Chronos: an error message is added to AutoformerForecaster.predict() (#5483)

When the input data to AutoformerForecaster.predict() is incorrect, the error message is confusing. Thus, a new error message is added.

* [Nano] Openvino quantization notebooks with nano (#5491)

* notebook with nano

* load with nano

* modify readme

* clear all outputs and enable benchmark_app

* fix typo

* Correct the path of start-fl-server.py in assembly.xml (#5509)

* add assembly

* Enable FL Sever in SGX

* Add start-fl-server.py to Docker Image

* rm ppml-assebly.xml

* Add start-python-fl-server-sgx.sh to Docker Image

* Put start-fl-server.py into assembly

Co-authored-by: Jiaming <[email protected]>

* fix compile issue for spark2.3 (#5501)

* Nano: update ipex_bf16_inference_model (#5470)

* rollback requirement-doc

* Update

* Update

* Update

* Update

* Fix ipex with amp error

* Fix unit test

* Update

* Remove redundant method

* Fix error

* remove redundant import

* Update ut to conver cases

* Fix the problem of cannot find the directory. (#5518)

* add assembly

* Enable FL Sever in SGX

* Add start-fl-server.py to Docker Image

* rm ppml-assebly.xml

* Add start-python-fl-server-sgx.sh to Docker Image

* Put start-fl-server.py into assembly

* Fix the target directory

Co-authored-by: Jiaming <[email protected]>

* [Orca] Replace pyarrow dependency in tf2 ray estimator (#5465)

* remove pyarrow dependency in tf2 ray backend

* update ut to compare prediction results

* CICD: re-organize existing workflows into two workflows: nightly_build and nightly_test (#5511)

* reorg workflows
* remove duplicate yamls
* add dependencies

* PPML add pytorch incremental training and tutorial (#5379)

* Orca: add save and load model doc to PyTorch Estimator quick-start. (#5504)

* feat: add save and load model doc to pytorch estimator quickstart.

* fix: fix typo.

* fix: fix typo

* fix: fix typo.

* feat: add doc to ray backend quickstart

* fix: fix typo.

* feat: add doc to ray backend quickstart

* Orca: add shutdown API to tf2 pyspark estimator. (#5499)

* feat: add shutdown API to tf2 pyspark estimator.

* fix: replace raise with log4Error.

* refactor: refactor the stop log server function.

* fix: fix code style issue.

* update nightly Scala build (#5523)

* Orca: refactor HDFS operation in estimator (#5325)

* fix: use default value if no hdfs port specified.

* fix: add default hdfs port when saving pkl

* refactor: replace pyarrow deps in estimator.

* fix: add process wait()

* fix: resolve unnecessary change

* Update nano training ipex bf16 (#5333)

* Update

* Fix code style

* re-run action

* Fix code style

* re-run action

* Update

* Fix code style

* support bf16 multi training

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* reduce ut time and re-run action

* track avx512

* Update lite bf16 training

* Update

* Update

* Update bf16 api

* Update

* Fix typo

* Orca: add get remote file to local decorator in orca file utils. (#5508)

* feat: add get remote file to local decorator in orca file utils.

* feat: add unit-test for local and s3 FS.

* fix: update func name and comments.

* fix: fix code style.

* fix: fix code style.

* fix: fix typo.

* fix: fix import.

* PPML FL add PSI support for fl nn server (#5434)

* [Orca] Refactor tf2 pyspark estimator save model (#5425)

* fix save model in pyspark est

* use a function to replace

* use orca.data to replace dllib

* Add Xshards test to github actions (#5512)

* Add Xshards test to github actions

* add file path

* fix yml

* fix csv path

* fix csv path

* fix yml

* add wget

* fic typo

* fix typo

* change csv file path

* update pwd

* add new line

* delete a line

* add time

Co-authored-by: yansu <[email protected]>

* Update some packages' edition (#5527)

* add assembly

* Enable FL Sever in SGX

* Add start-fl-server.py to Docker Image

* rm ppml-assebly.xml

* Add start-python-fl-server-sgx.sh to Docker Image

* Put start-fl-server.py into assembly

* Change some packages' edition

* Fix the problem of fl/start-fl-server.py

Co-authored-by: Jiaming <[email protected]>

* Use tracker.py to build xgboost4j_2.12-1.1.2.jar (#5513)

Reason to change: tracker.py in ${BIGDL_HOME}/jars/bigdl-dllib-spark_${SPARK_VERSION}-${BIGDL_VERSION}.jar is not functional.

Instead of using tracker.py to build ${BIGDL_HOME}/jars/bigdl-dllib-spark_${SPARK_VERSION}-${BIGDL_VERSION}.jar,
we now use it to build xgboost4j_2.12-1.1.2.jar, which can be reached when executing xgboost examples.

* Create nightly-build-example-tests-ppml.yaml (#5517)

* Create nightly-build-example-tests-ppml.yaml

* Update nightly-build-example-tests-ppml.yaml

Triggered by schedule

* Update nightly-build-example-tests-ppml.yaml

* Chronos: add tune for lstm and s2s (#5521)

* set use_hpo=True and add ut

* PPML: add k8s to Gramine Docker (#5428)

* feat: add k8s related files

* fix: change secure-argv to secure_argv in Gramine

* fix: check if secured_argv exists with -f

* fix: exit when not ready in init.sh

* fix: change path of encalve-key in k8s template

* feat: copy k8s directory to Gramine directory

* mod: version in development

* mod: enclave-key path

* del: remove tmp build action

* feat: Simplify deployment operations for users

* doc: update README.md

* doc: correct a typo

* doc: correct typo

* add spark pi on k8s scripts

* ppml python psi hashing (#5531)

* cicd: refactor the github actions workflows (#5529)

* add composite actions

* add manually build

* Nano: Enhance optimizer pipeline (#5490)

* refactor code

* enable get_best_model

* basic impl finnish

* add license

* clean openvino output

* remove one line

* remove partial output of openvino and inc

* add trace and quantize

* update based on comments

* ingore errors

* update based on comment

* update based on comment

* add type hint for return

* fix issues (#5539)

* Chronos: set kernel of Autoformer to odd number (#5534)

* set kernel to odd

* [PPML] Change Occlum tmp_dir  to use hostfs (#5520)

* change_tmp_dir_hostfs，but xgboost example will crash (Need to find the reason).

* add a switch to use tmp hostfs

* [PPML] Implement verifyQuote in QuoteVerifier.scala (#5495)

* Add SGXDCAMQuoteVerifierImpl in for QuoteVerifier
* Add unit test for SGXDCAPQuoteVerifierImpl
* Move attestation.Attestation to dcap.Attestation

* [ORCA] Add xshards example (#5542)

* update sh

* add data example 2

* add data example 2

* fix style

* fix time

* fix time

* fix time

Co-authored-by: yansu <[email protected]>

* add psi in nn tutorial (#5532)

* [PPML] refine eHSN-KMS according to user experience (#5544)

* Add python hint to bigdl.friesian (#5430)

* friesian.feature.table

* script refactor

* doc

* [PPML] Fix XGBoost example in docker-graphene/README.md (#5510)

* Fix the error in docker-graphene/README.md
* Add instructions on how to process Boston_Housing.csv
* Change to use the correct path for xgboost examples

* [ORCA] Add Xshrads tutorial test (#5548)

* add tutorial test

* fix typo

* fix name

* remove shards_regressionExample.py

* update yml

* update yml

* update yml

* update yml

* update yml

* add readme

* update yml

* update yml

Co-authored-by: yansu <[email protected]>

* Change the environment settings in Dockerfile for FL Server. (#5554)

* add assembly

* Enable FL Sever in SGX

* Add start-fl-server.py to Docker Image

* rm ppml-assebly.xml

* Add start-python-fl-server-sgx.sh to Docker Image

* Put start-fl-server.py into assembly

* Improve the Dockerfile to create right environment for FL Server

* Fix the problem of pytorch-nn-lr-tutorial.md

Co-authored-by: Jiaming <[email protected]>

* Reduce XShards tutorial test time (#5555)

* Update README of XShards tutorials (#5557)

* update readme

* minor

* add orca ray python test to github action (#5550)

* add orca ray python test to github action

* update

* update

* update

* update

* update

* update

* [PPML] Remove sgxsdk quote verify in ut (#5553)

* Remove sgxsdk quote verify in ut

* Use Env SGXSDK for UT

* Refine

* [PPML] Refine get RABIT_TRACKER_IP env in XGBoost examples (#5528)

* The original code will lead to keyError if "RABIT_TRACKER_IP" is not
set. The ideal result should be return None.

* Change the Dockerfile for better. (#5559)

* add assembly

* Enable FL Sever in SGX

* Add start-fl-server.py to Docker Image

* rm ppml-assebly.xml

* Add start-python-fl-server-sgx.sh to Docker Image

* Put start-fl-server.py into assembly

* Change the Dockerfile for better

Co-authored-by: Jiaming <[email protected]>

* cicd: add a new workflow called PR_Validation (#5546)

* add PR_Validation workflow

* test

* test

* test

* test

* test

* test

* test

* add friesian ut to PR validation

* revert changes

* add dllib ut

* test

* revert

* [PPML] refine EHSM-KMS and modify annotations (#5563)

* add spark sql examples (#5535)

* add spark sql local mode scripts

* update local ip value

Co-authored-by: baishaojie <[email protected]>

* fix prometheus-client version for ray dashboard (#5549)

* Fix bug that string_embed function Only replace one columns when replace=True. (#5497)

* add for loop to replace all Columns.

* add replace test for string_embds function

* Orca: Add orca ray ctx example to run on self hosted (#5484)

* Add orca ray ctx example in github workflows

* Install cmake 3.16.3 with pip

* Increase the CPU limits

* Add newline at EOF

* Align with the new format

* Add the pull_request trigger

* Add missing property

* Uninstall spark first

* Uninstall spark3

* Remove test triggers

* Add tests to PR_validation

* Fix a bug

* Test the validation trigger

* Undo the change

* Chronos: Add `from_tsdataset` method, BaseTF2Forecaster can input a tsdataset (#5064)

* add from_tsdataset for tfForecaster

* fix known issues

* fix known issues again

* Add attestation guide to README.md in ppml/docker-graphene (#5568)

* Add attestation guide to README.md in ppml/docker-graphene

* Refine

* Refine

* Refine

* Refine

* Refine

* Remove hard coded spark mode in xgboost examples (#5579)

* kms-utils README fix (#5582)

* rename inference ppl test to move to tests-ipex (#5581)

* [PPML] Fix ppml code style (#5584)

* Add BigDL-core dep and scope
* Fix ppml code style

* Chronos : add [how to tune a forecaster model] notebook (#5399)

* add how_to_tune_a_forecaster_model notebook

* modify step1: data preparation

* update based how to train

* modify a bug of summary

* update based on comment

* move to howto

* fix typo

* add file

* update based on comments

* rename the tile

* Fixprocessing for several orca tutorials so it can use the original dataset download from website (#5572)

* [PPML] Modify FTP_URI for sgxdcap quote verifier ut (#5591)

* Modify FTP_URI for sgxdcap quote verifier ut

* Fix

* Fix

* Add shutdown api (#5592)

* feat: add shutdown api

* test: add unit test for shutdown API.

* add orca py37 UT to github action (#5578)

* add orca py37 UT to github action

* update pull_request path

* update build python whl

* update bigdl-core version

* remove pull_request

* remove scala app test for nightly test (#5595)

* Fix yaml load (#5597)

* PPML FL FGBoost save/load server model (#5577)

* remove hardcode credentials (#5598)

* PPML FL protobuf and generated update (#5604)

* PPML FL Pytorch NN tutorial refine (#5545)

* Update powered-by.md

* Update powered-by.md

* optim get schema and get class name in shards into 1 rdd operation (#5573)

* update output file to hdfs path (#5593)

* Chronos: Remove UTs for all `Deprecated` API and `tf1` (#5530)

* delete deprecated UTs

* remove tf1 UTs

* Add spark auth secret env to occlum (#5602)

* Add Azure KMS support for PPML python api (#5547)

* add azure kms for python

* update ppml doc

* update ppml tutorial readme

* update kms-utils entrypoint.sh

* update entrypoint.sh

Co-authored-by: Zhou <[email protected]>

* [Orca] Refactor `model_dir` as an option in tf2 pyspark estimator (#5541)

* refactor model_dir as an option

* modify ut to test non model_dir

* update coding format

* update return results

* add support in load api

* move model_dir from ut

* Pytorch VFL NN refactor, refine and example test (#5607)

* [Documentation] Add Extension nbsphinx-link (#5596)

* Add extension nbsphinx-link: link notebook out of the sphinx root folder to the sphinx project

* Link Notebook custom.ipynb directly from python folder to the sphinx project

* Link Notebook seq_and_func.ipynb directly from python folder to the sphinx project

* Small title fix

* Make interactive graphs visible in readthedocs webpage for seq_and_func.ipynb

* Regenerate outputs for custom.ipynb and remove parts of them

* add five components to tpp (#5614)

Co-authored-by: Jin Hanyu <[email protected]>

* Chronos: change get_latency to a static method (#5618)

* Implement get quote from ehsm (#5621)

* Implement getQuoteFromServer in EHSMAttestationService

* Fix

* Remove Base64 lib

Co-authored-by: xiangyuT <[email protected]>

* Chronos : add regression functions for forecast metrics to replace torchmetrics (#5402)

* add regression functions

* delete unexpected import

* add function api

* remove eval function

* revised the problems mentioned

* Modify files imported TORCHMETRICS_REGRESSION_MAP

* fix bug caused by numpy function

* fix pep8

Co-authored-by: Xinyi Zhang <[email protected]>
Co-authored-by: theaperdeng <[email protected]>

* Fix Friesian READMEs (#5626)

* Chronos: installation panel enhancement (#5588)

* Update:installation panel enhancement
Left a bug to fix and install CMD to revise.

* Bugfix: Fixed a bug when using Firefox browser to use the install guide form.
Fixed the bug that the table's background color may be changed by mistake on Firefox browser.

* Update: update panel as comment required.

* Bugfix: fixed a bug in js file that will show wrong cmd when clicking 'windows' or 'docker'.

* Bugfix: fixed the bug that 'cmd' shows incorrectly when clicking 'prophet' and 'ARIMA'.

* support convert rdd of pandas df to spark df with arrow (#5385)

* Doc: update hadoop document. (#5623)

* doc: update hadoop document.

* fix: fix wording.

* fix: fix wording.

* fix: fix wording.

* Nano: Add pytorch-lightning bf16 training example (#5526)

* rollback requirement-doc

* Add Nano BF16 Training example

* Update

* Update

* Chronos: numpy and python version in pr validation Github Action is fixed (#5613)

* [Nano] Nano How-to Guides: Format & PyTorch Inference (#5480)

* Create doc tree index for Nano How-to Guides

* Add How to guide for PyTorch Inference using ONNXRuntime

* Add How to guide for PyTorch Inference using OpenVINO

* Update How to guide for PyTorch Inference using OpenVINO/ONNXRuntime

* Change current notebook to md and revise contents to be more concentrated

* Add How-to Guide: Install BigDL-Nano in Google Colab (need further update)

* Revise words in How-to Guide for PyTorch Inference using OpenVINO/ONNXRuntime

* Add How-To Guide: Quantize PyTorch Model for Inference using Intel Neural Compressor

* Add How-To Guide: Quantize PyTorch Model for Inference using Post-training Quantization Tools

* Add API doc links and small revision

* Test: syncronization through marks in py files

* Test: syncronization through notebook with cells hidden from rendering in doc

* Remove test commits for runnable example <-> guides synchronization

* Enable rendering notebook from location out of sphinx source root

* Update guide "How to accelerate a PyTorch inference pipeline through OpenVINO" to notebook under python folder

* Update guide "How to quantize your PyTorch model for inference using Intel Neural Compressor" to notebook under python folder

* Fix bug that markdown will be ignored inside html tags for nbconvert, and notebook revise

* Update guide 'How to quantize your PyTorch model for inference using Post-training Optimization Tools' to notebook under python folder

* Small updates to index and current guides

* Revision based on Junwei's comments

* Update how-to guides: How to install BigDL-Nano in Google Colab, and update index page

* Small typo fix

* Fix kms-utils README

* Fix enroll app command in README

* split assembly (#5564)

* split assembly

* update

* update

* [PPML] Refine occlum readme and config (#5625)

* Refine occlum readme and config

* Update docker sql config

* Reduce instance to 1

* [PPML] Add user doc for how to install python dependencies (#5494)

* Add user doc in docker-graphene/README.md
* Add instructions on how to add egg dependencies

* [PPML] Fix file_name too long issue (#5628)

Test xgboost and tpch example successfully.

* [PPML] Refine xgboost docker example config (#5631)

* Chronos : Penalize underestimation example (#5627)

* add example

* remove valid dataset

* fix typo

* modify parameters and graph

* update based on comment

* modify index link

* modify js

* update

* Chronos: Sync SparkXShards changes(TCMFForecaster) (#5630)

* Fix xshards tensorflow (#5633)

* fix

* fix

* fix

* fix

* Chronos: Delete existing tags and replace with new Installation options (#5165)

* remove skipif tf2

* delete skip_onnxrt tags

* fix synatx error

* add installation options

* remove invalid options

* fix known issues

* fix known issues

* rollback tf1 uts

* rollback tf1 UTs

* rollback invalid changes

Co-authored-by: fuckyouNoob <[email protected]>

* Force update bigdl-core jar during build (#5569)

* Add -U in dllib release
* Add -U to orca and friesian

* Chronos: make `roll=True` the default value for `to_torch_data_loader` (#5605)

* Chronos: modify examples of `.evaluate` (#5622)

* fix nano UT (#5650)

* [PPML]Remove simple KMS entroll command (#5624)

* fix: simple enroll

* Update DeepRec README.md (#5652)

* fix dllib pom scope and typo in examples (#5575)

* fix dlliv typo in examples

* [PPML] Fix bash expected unary operator (#5653)

* [PPML] Implement bi-attestation in AttestationCLI.scala (#5648)

* Implement bi-attestation in AttestationCLI.scala

* Refine

* Refine

* Refine

* Refine

* Fix problems according to comments

Co-authored-by: xiangyuT <[email protected]>

* PPML FL change clientID from String to Int (#5647)

* support convert spark df to shards with arrow (#5558)

* Chronos: syntax error in Evaluator API documentation is fixed (#5660)

* Chronos: how to create a forecaster (#5172)

* how to guide from_tsdataset

* improve md

* remove redundant message

* fix known issues

* add rst

* move create-forecaster to new line

* fix known issue

* fix syntax error

* fix syntax error again

* add open-colab icon

* fix known issues

* fix some words

* fix typo

* fix syntax error

* add some update

Co-authored-by: theaperdeng <[email protected]>

* Nano : ResNet Demo for InferenceOptmizer (#5580)

* add resnet demo

* add basic readme

* add inference result

* update based on comment

* update based on comment

* fix style

* fix typos and update based on comment

* update numpy version

* update for faster demo

* modify some number

* [PPML] Add SGX plugin device readme (#5663)

* add sgx plugin device readme

* fix tutorial xshards env (#5661)

* fix tutorial xshards env

* add protobuf

* add torch

* add pyarrow

* add python requirements

* Nano : fix and enhancement of InferenceOptimizer (#5611)

* fix accuracy-related bug and add model.eval()

* automatically print optimization results

* dynamic control latency_sample_num

* add ut for no optimize

* fix accuracy

* update ut

* update based on comment

* remove statue restore

* Nano: Update Nano PyTorch Lightning tutorial test script (#5612)

* Update

* Add test for lightning bf16

* Update

* Fix shell

* remove redundant code in resnet18 example (#5665)

* support init_orca_context with spark.master=local[*, F]/local[N, F] (#5651)

* support format local[N,F] & local[*,F]

* add comment and UT

Co-authored-by: Zhou <[email protected]>

* Chronos: how to speedup inference on one node (onnx, openvino) (#5556)

* how to speedup inference on one node

* index added

* fixed

* fixed

* split to 2 guides

* add build_onnx and build_openvino

* add note border

* fix

* fix colab conflict

Co-authored-by: binbin <[email protected]>

* Nano: Add nano pytorch example (#5570)

* rollback requirement-doc

* Add example

* Update

* Update

* Update

* Update

* Add comments and remove testing loop

* Add comments about linear scale and warmup

* remove .keep file

* add create-job-status-badge action (#5676)

* [Nano] add __call__ method for OpenVINOModel (#5540)

* [Nano] support __call__ method for OpenVINOModel

* remove unnecessary kwargs in __call__ method of OpenVINOModel

* specify which __call__ method to call in KerasOpenVINOModel

* remove whitespace

* make KerasOpenVINOModel and PytorchOpenVINOModel not inherit from OpenVINOModel

* fix docstring of OpenVINOModel._save_model

* remove extra import

* fix _save_model

* change_save_model

* simplify _model_exists_or_error

* Support decimal for Spark DataFrame (#5671)

* support decimal

* remove

* Fix yaml load (#5655)

* [PPML] Fix EHSM 0.3.0 create enclave failure (#5683)

Co-authored-by: glorysdj <[email protected]>

* fix issue 4642, fix DBFS file path problem on Dataricks (#5679)

* fix issue 4642

* parse model_dir

Co-authored-by: Zhou <[email protected]>

* Change Analytics Zoo related to BigDL in doc (#5685)

* change az to bigdl

* minor

* minor

* Add Ray arguments in init_orca_context to Doc page (#5686)

* add ray docs

* fix

* fix

* fix style

* minor

* scala lightGBM classifier and regressor[WIP] (#5447)

* scala lightGBM classifier and regressor

* Fix Orca Example styles (#5692)

* part1

* part2

* part3

* part4

* exclude and finish

* Chronos: fix spark 3.1 bug in xshards unscale (#5689)

* Add selfHostedRunner input (#5701)

* add create-job-status-badge action

* add selfHostedRunner input

* fix: update redis port random range. (#5702)

* [Nano] Add Github Action Tests for How-To Guides: Inference PyTorch (#5668)

* Disable other nano action tests temporarily

* Add github action tests for Nano How-To Guides: Inference > PyTorch

* Add dependency install for pytest and nbmake

* Separate tests based on openvino/onnx dependencies

* Tighten the scope for sed to change

* Fix [AttributeError: 'OutStream' object has no attribute 'watch_fd_thread'] when testing for OpenVINO related notebook
Downgrade ipykernal to solve this

* Edit trigger conditions for how-to gudies tests

* Enable other nano tests again

* Style fix: remove extra tabs and add missing comments

* Add how-to guides tests trigger limitations for push on main branch

* fix integration test for dllib examples (#5707)

* Chronos: train tcn model on gpu and speed up inference on cpu (#5594)

* add nano_gpu

* simplify the code

* simplify the code

* modify the code

* add some updates

* add document

Co-authored-by: theaperdeng <[email protected]>

* bigdl all in one jar (#5703)

* bigdl all in one jar

* update

* cicd: scan dependency tree to help debug if fossa scan failed (#5698)

* scan dependency

* if always

* update

* filter dependency tree if failure

* revert

* [PPML] EHSM 0.3.0 AS Support for Gramine (#5644)

* [PPML]correct the command and add explanation in kms-utils README (#5664)

* fix command argument in README

* explain file structure

* typo

* diff between file and directory

* typo

* commeng location

* make comment and variable understandable

* fix issue-702 (#5708)

* Change the environment of container and fix the problem of cannot compare two variables. (#5620)

* add assembly

* Enable FL Sever in SGX

* Add start-fl-server.py to Docker Image

* rm ppml-assebly.xml

* Add start-python-fl-server-sgx.sh to Docker Image

* Put start-fl-server.py into assembly

* Change the environment of container and fix the problem of cannot compare two variables

* Change the location of command "rm protobuf" in Dockerfile

* Add explicit cast to avoid thread being syspended.

* Change docker-gramine/Dockerfile to fit FL's environment requirements.

Co-authored-by: Jiaming <[email protected]>

* Chronos: PR validation-spark3.1-ray job to Github Action (#5681)

* pytest ray-part1

* pytest ray-part1+part2

Co-authored-by: binbin <[email protected]>

* Add Friesian README (#5709)

* Create README.md

* Update README.md

* Update README.md

* Add create dynamic badge job (#5706)

* add create-job-status-badge action

* add selfHostedRunner input

* add create badge job

* change isSelfHostedRunner to is-self-hosted-runner

* Nano : migrate trace and quantize from Trainer to InferenceOptimizer (#5667)

* basic migration

* add load_xxx_model

* modify init

* Chronos: install openvino-dev in Github action (#5704)

* install openvino

* fix

* fix

* numpy==1.19.5

Co-authored-by: binbin <[email protected]>

* Chronos: Fix `LazyImport` known issue (#5502)

* fix known issues

* fix known isses

* Add nano tensorflow examples (#5463)

* add tensorflow multi instance training
* add tensorflow sparse embedding example

* Chronos: Add support for `id_sensitive`=True to Forecaster.from_tsdataset (#5551)

* supports id_sensitive=True

* new params is_predict

* rollback is_predict

* fix known issues

* fix code style

* Chronos: how to guide page cleaning (#5657)

* update

* add nblink

* windows guide fix

* add some updated

* update chronos

* add update image

* add new image

* add updates

* add update

* add new image

* fix typo

* fix notebook location

* add new index

* update

* fix some doc

* update tree

* fix typo

* Fix the problem of indents. (#5714)

* add assembly

* Enable FL Sever in SGX

* Add start-fl-server.py to Docker Image

* rm ppml-assebly.xml

* Add start-python-fl-server-sgx.sh to Docker Image

* Put start-fl-server.py into assembly

* Fix the problem of indents.

Co-authored-by: Jiaming <[email protected]>

* Change the editions of torch and torchvision. (#5717)

* add assembly

* Enable FL Sever in SGX

* Add start-fl-server.py to Docker Image

* rm ppml-assebly.xml

* Add start-python-fl-server-sgx.sh to Docker Image

* Put start-fl-server.py into assembly

* Change the editions of torch and torchvision

Co-authored-by: Jiaming <[email protected]>

* Pentest Issue: remove hardcode credentials (#5649)

* remove hardcode credentials

* update

* add orca ut to PR_validation.yml

* use redis_paasword in extra_params otherwise use the default password in ray

* refine code

* update

* typo

* remove orca ut from pr validation

* update

* update

* Modify two errors in create-job-badge and license-scan workflow. (#5718)

* add create-job-status-badge action

* add selfHostedRunner input

* add create badge job

* change isSelfHostedRunner to is-self-hosted-runner

* update create-job-status-badge

* adjust indent

* Update create aks script on Graphene (#5566)

* update create aks

* update

* update

* fix create sp

* fix

* make executable for az scripts

* install command

Co-authored-by: root <root@jiao-test.rfprax1kp2xefkkygl2c02qm5g.cx.internal.cloudapp.net>

* Add long descriptions to pypi releases (#5711)

* long desc

* style

* meet comment

* Update PR_validation.yml (#5722)

* add orca horovod tf2 test to github actions (#5601)

* add orca horovod tf2 test to github actions

* update pull_request

* update

* add keras

* update python version

* update horovod

* update cmake

* install bigdl-orca-spark3

* update

* update mxnet

* update

* update numpy

* update

* update ray version

* unintsall intel-tensorflow

* add conda env

* install horovod

* remove conda env

* install bigdl-orca-spark3

* split conda env

* source conda

* split conda env

* update

* update env

* remove env

* update

* update

* uninstall requirements

* update

* update install whl

* update

* add remove env

* install bigdl-orca-spark3

* update

* Add conditional judgment in scala-style-check and python-style-check (#5721)

* add create-job-status-badge action

* add selfHostedRunner input

* add create badge job

* change isSelfHostedRunner to is-self-hosted-runner

* update create-job-status-badge

* adjust indent

* update python-style-check and scala-style-check

* Update scala-style-check.yml

* Update python-style-check.yml

* [PPML] Initiate README.md for AttestationCLI (#5688)

* Init

* Refine

* Refine

* Refine according to comments

* Chronos: install openvino in Chronos-PRVN-Spark3.1-ray check (#5719)

Co-authored-by: binbin <[email protected]>

* Chronos: PR validation-spark3.1-onnxrt job to Github Action (#5725)

Co-authored-by: binbin <[email protected]>

* Chronos: test onnxrt==1.11.1 in PR validation-spark2.4-onnxrt job (#5677)

* onnxrt==1.11.1

* fix

Co-authored-by: binbin <[email protected]>

* [PPML] change KMS appkey to apikey (#5658)

* change appkey to apikey

* change APiKEY to APIKEY

Co-authored-by: Heyang Sun <[email protected]>

* [PPML]delete unused argument in kms-utils README(#5733)

* [PPML] Fix format error of quote from attestation service (#5732)

* Fix format error of quote fom attestatioon service

* Refine

* Refine

* skip dummy verify

Co-authored-by: xiangyuT <[email protected]>

* release fat jars (#5736)

* release fat jars

* format

* [DLlib] GBT CriteoClickLogsDataset example (#5723)

* init gbt class

* remove something unimportant

* add readme

* change xgb to gbt

* use overwrite to save

* add text to fix message=Header does not match expected text line=1

* Chronos: 'forecaster.predict_with_openvino' supports quantization API (#5690)

* add openvino quantization

* add some comments

* fix code style

* update assembly pom (#5737)

Co-authored-by: Zhou <[email protected]>

* Add more data types in convert DataFrame to numpy (#5680)

* add more datatype

* update unit test

* update ut in test_estimator_pyspark_backend

* update ut in test_estimator_pyspark_backend

* update ut in test_estimator_ray_backend

* update ut in test_estimator_ray_backend

* add array decimal type

* update array data type

Co-authored-by: Zhou <[email protected]>

* copy config (#5731)

* Nano : Enhancement for output format of InferenceOptimizer (#5705)

* first commit

* improve output of optimize

* fix method_type for method lack of dependency

* update openvino model for thread num

* add thread num for trace and quantize

* update based on comment: modify output, add progress bar and remove method type

* modify latency to latency(ms)

* add thread_num for Trainer.trace and Trainer.quantize

* fix bug of openvino cpu_num

* [PPML]doc: change appkey to apikey in kms README (#5747)

* cicd: add debug info to license-scan workflow (#5739)

* add debug info

* debug

* echo

* update

* update

* Add nightly-test job badge (#5750)

* add create-job-status-badge action

* add selfHostedRunner input

* add create badge job

* change isSelfHostedRunner to is-self-hosted-runner

* update create-job-status-badge

* adjust indent

* update python-style-check and scala-style-check

* Update scala-style-check.yml

* Update python-style-check.yml

* Update nightly_test.yml

* refactor assmbly deploy (#5753)

* Fix702 (#5727)

add lightgbm scope

* [PPML] Implement VerificationCLI to verify Attestation Service (#5742)

* Implement VerificationCLI and a bash to use VerificationCLI

* Refine

Co-authored-by: xiangyuT <[email protected]>

* Add ipython extension for rendering code blocks in excuted notebooks properly in our readthedocs project (#5757)

* [Orca] Refacor `model_dir` as optional in pytorch pyspark estimator (#5738)

* refactor model_dir as optional in pytorch estimator

* update ut to test non model_dir

* support standard scaler for shards (#5716)

* support standard scaler

* support param map for lightGBM (#5712)

* Support load state dict in TorchNano (#5687)

* support load state dict in TorchNano
* fix UT

* add http prometheus client (#5743)

* [PPML] Add occlum gbt readme and fix file path (#5756)

* change path examples to example

* add run_spark_gbt_criteo.sh

* update readme and timeout config

* some improvement (#5765)

* add conda env to orca test actions (#5735)

* add conda env to orca test actions

* add pull_request

* update env

* update

* update

* update

* update

* update

* update cmake

* update ut

* update ut

* update requirements horovod pytorch

* update

* update

* update protobuf

* update horovod test

* add tensorflow

* update tensorflow

* update

* update python exampletet

* update python UT

* update

* update

* update

* [PPML] Remove XGBoost from PPML guide

* Nano : reduce time cost for InferenceOptimizer and update demo (#5740)

* update readme and reduce time cost for calculating accuracy

* add prune type

* filter warnings

* add check for input_sample

* update based on comment

* update readme and add ut

* update input_sample

* delete redundant lines

* add forward_args

* Chronos: add conda env to Chronos PRVN Actions (#5744)

* [PPML] Fix gbt example spark config error (#5774)

* Fix the conflict of `KMP_AFFINITY` and onnxruntime  (#5764)

* remove `compact` option from KMP_AFFINITY's default value

* [PPML] Fix tracker bind issue happens in xgboost examples with SGX enabled (#5766)

* Change tracker.py to acquire IP automatically

* Remove the dependencies in documents on RABIT_TRACKER_IP

* Set xgboost.spark.ignoreSsl to True in xgboost examples

* Add deprecate hint for xgboost examples

* [PPML]Add log file argument to bigdl-ppml-submit.sh (#5726)

* add log file argument

* change demo for log argument

* Orca OpenVINO estimator add input order param & dataframe input support multi-output (#5471)

* update

* df support multi-output

* temp

* update

* update test

* meet comments

* meet comments

* del pred

* update testcase

* [Nano] How-To Guides: Training - PyTorch Lightning (#5643)

* Rearrange file structure for PyTorch Inference for docs and add titles for PyTorch-Lightning Training

* Add How-to guide: How to accelerate a PyTorch-Lightning application on training workloads through Intel® Extension for PyTorch*

* Add how-to guide: How to accelerate a PyTorch-Lightning application on training workloads through multiple instances

* Revise: remove '-' in 'PyTorch-Lightning' and some other changes

* Add How-to guides: How to use the channels last memory format in your PyTorch Lightning application for training

* Add how-to guide: Use BFloat16 Mixed Precision for PyTorch Lightning Training

* Add How-to guide: How to accelerate a computer vision data processing pipeline

* Small Revision: change comments in several code cells

* Disable other nano tests temporarily

* Add github action tests for Nano Training PyTorch Lightning tests

* Enable other nano tests again

* Small revisions: typos and explanation texts changes

* Revise: update based on comments

* Chronos: fix `quantize` and `to_torch_data_loader` API doc (#5755)

* Add region proxy for create-job-badge action (#5761)

* add create-job-status-badge action

* add selfHostedRunner input

* add create badge job

* change isSelfHostedRunner to is-self-hosted-runner

* update create-job-status-badge

* adjust indent

* update python-style-check and scala-style-check

* update

* Update scala-style-check.yml

* Update python-style-check.yml

* add nightly-test job badge

* Update nightly_test.yml

* Add region proxy

* Chronos: fix install cmd in Chronos User Guide document (#5776)

* PPML: Initiate doc for verificationCLI (#5770)

* Initiate doc for verificationCLI and refine

* Update README.md

Co-authored-by: xiangyuT <[email protected]>

* PPML: mod TPCH document (#5583)

* doc: modify markdown grammar typo

* doc: md typo

* typo

* fix

* fix

* fix

* hint

* correct argument

* add path of password to container

* fix syntax errors

* fix

* fix

* [Nano] Update Trainer -> InferenceOptimizer for Related How-to Guides (#5778)

* Update Trainer to InferenOptimizer for related how-to guides

* Disable other nano tests temporarily

* Enable other nano tests again

* tdx/scripts (#5636)

* Fix runnerHostedOn input (#5784)

* add create-job-status-badge action

* add selfHostedRunner input

* add create badge job

* change isSelfHostedRunner to is-self-hosted-runner

* update create-job-status-badge

* adjust indent

* update python-style-check and scala-style-check

* update

* Update scala-style-check.yml

* Update python-style-check.yml

* add nightly-test job badge

* Update nightly_test.yml

* Add region proxy

* fix runnerHostedOn input

* tdx/docker/bigdl-tdx (#5637)

* tdx/docker/bigdl-tdx

* address comment

* Lightgbmpython (#5696)

Port scala lightgbm to python

* [Orca] Support save model weights in tf2 ray estimator (#5780)

* add save load weights func

* hot fix pyspark bug

* add ut

* code style fix

* Nano : add API for InferenceOptimizer and deprecated some Trainer APIs (#5768)

* add api and deprecated

* add license

* add deprecated

* update comment for api display

* modify comment

* update based on comment

* [Nano] Update Trainer -> InferenceOptimizer for Related Examples (#5781)

* Update Trainer to InferenceOptimizer for related examples

* Diasble unrelated nano tests

* Enable unrelated nano tests again

* Doc: add ray port FAQ to ray overview document. (#5752)

* feat: add faq to ray overview doc.

* fix: fix doc style.

* fix: fix typo.

* feat: add ray redis and dashboard port conflict to faq.

* fix: fix redis port param name.

* fix: fix redis port param name.

* fix: fix typo

* fix: fix typo

* fix: fix wording.

* tdx/docker/client-image (#5639)

* Add Friesian multi task example code and readme (#5460)

* add multi task example codes and readme

* fix for Friesian code style

* part 1: fix comments in pr

* update for comments part 2

* update for pr comments part 3

* update for code style

* update codes and add test results

* Update README.md

* Add Nano openvino async api (#5615)

* add openvino async_predict api

* add unit test for async_predict api

* fix codestyle error

* optimize input/output format and doc string

* change para name, docstring & add type hint

* fix error when calling on_forward_end after the output format of __call__ in OpenVINOModel is changed

* fix codestyle check error & disable mypy type check

* fix codestyle check error

* use model_exist_or_error

* Fix orca.pandas.read_csv() only return top 10k rows issue (#5775)

* bug fix

* add ut

* add test file (#5788)

* update pod template (#5786)

* Add nano pytorch training bf16 example (#5600)

* rollback requirement-doc

* Add Nano PyTorch BF16 Example

* Add comments and remove testing loop

* Update

* Nano: Add tensorflow quantization example. (#5619)

* Nano: Add tensorflow quantization example.

* simplify quantize arguments

* add some extra comments

* delete concrete numbers

* change dataset to cifar10

* change file name to cifar10

* improve code organization

* remove unnecessary comments

* Chronos: how to evaluate using forecaster (#5710)

* [PPML] Change E2E XGB example to Spark GBT (#5785)

* change to xgb to gbt

* use overwrite to save

* Update multi-task example (#5787)

* update multi-task

* fix

* minor

* minor

* [PPML] Fix BigDL Graphene Azure scripts (#5767)

* update

* update submit scripts

* Chronos: fix input feature num != output feature num case for autoformer (#5673)

* fix input feature num != output feature num case

* add updates

* Lightgbm (#5789)

* error message

* update Databricks user guide (#5779)

* update databricks doc

* update databricks doc

* update databricks doc

* update databricks doc

* update databricks doc

* update databricks doc

Co-authored-by: Zhou <[email protected]>

* [PPML] Update end to end GBT example readme (#5792)

* update end to end gbt readme

* update header location

* add bigdl-ppml jar with denpendency jar in tdx docker (#5793)

* Fix PanTest issue（partly） -- ObjectInputStream's readObject method may deserialize untrusted data (#5666)

* fix readObject deserialize untrusted data

* use classTag[T] and use classof[modue]

* remove double init inputStream

* add accept pattern

* accept Array

* add EncodeUtils ut

* add success case and refine fail case

* add headers

* Add warning for Distributed XGBoost examples (#5795)

* tdx e2e readme (#5640)

* update graphene ppml examples to main (#5804)

* update graphene ppml examples

* update graphene ppml examples

Co-authored-by: Jin Hanyu <[email protected]>

* Fix a bug of readthedocs about get_public_dataset (#5808)

* Chronos: fix ZooTestCase in tensorflow unit tests (#5800)

* Chronos: decouple mtnet forecaster with orca (#5815)

* update mtnet model to decouple with orca

* rename one of the method

* update test-suites scripts (#5823)

Co-authored-by: Jin Hanyu <[email protected]>

* update tpp blas component information (#5825)

Co-authored-by: Jin Hanyu <[email protected]>

* Chronos : support confidence interval for forecaster (#5682)

* first commit, add predict_interval

* add return value of calculate

* modify parameter name, only calculate data noise once

* add ut for lstm, nbeats, s2s, tcn

* add paper link, change evaluate to predict

* modify data noise

* add ut for loader c and change to mean_yhat

* support confidence interval for autoformer

* update ppml examples scripts (#5827)

Co-authored-by: Jin Hanyu <[email protected]>

* Chronos/Nano: update nano and chronos action trigger domain (#5809)

* update nano and chronos action trigger domain

* update

* delete orca.learn as a trigger

* Support tf2.1 load model for ray backend (#5830)

* support tf2.1

* meet review

* Chronos: More autoformer bug fix and enhancement (#5741)

* midterm push for working in cafe:)

* add updates

* remove confusing comments

* consider the conflict between tsdataset.label_len and label_len

* update according to comments, fix pep8

* add test cmt

Co-authored-by: Jin Hanyu <[email protected]>
Co-authored-by: Jiaming <[email protected]>
Co-authored-by: Mingzhi Hu <[email protected]>
Co-authored-by: Zhentao Chen <[email protected]>
Co-authored-by: Cengguang Zhang <[email protected]>
Co-authored-by: Juntao_Luo <[email protected]>
Co-authored-by: binbin Deng <[email protected]>
Co-authored-by: dding3 <[email protected]>
Co-authored-by: SONG Ge <[email protected]>
Co-authored-by: Shaojun Liu <[email protected]>
Co-authored-by: Nancy <[email protected]>
Co-authored-by: yansu <[email protected]>
Co-authored-by: Guancheng Fu <[email protected]>
Co-authored-by: ZehuaCao <[email protected]>
Co-authored-by: Ruonan Wang <[email protected]>
Co-authored-by: Shaojie Cui <[email protected]>
Co-authored-by: Wang Jian <[email protected]>
Co-authored-by: Xiangyu Tian <[email protected]>
Co-authored-by: Heyang Sun <[email protected]>
Co-authored-by: Zhao Changmin <[email protected]>
Co-authored-by: Kai Huang <[email protected]>
Co-authored-by: pinggao187 <[email protected]>
Co-authored-by: shaojie <[email protected]>
Co-authored-by: baishaojie <[email protected]>
Co-authored-by: Shan Yu <[email protected]>
Co-authored-by: Aixi Deng <[email protected]>
Co-authored-by: liangs6212 <[email protected]>
Co-authored-by: Qiyuan Gong <[email protected]>
Co-authored-by: Jason Dai <[email protected]>
Co-authored-by: Jiao Wang <[email protected]>
Co-authored-by: Jian Zhou <[email protected]>
Co-authored-by: Zhou <[email protected]>
Co-authored-by: Yuwen Hu <[email protected]>
Co-authored-by: Jin Hanyu <[email protected]>
Co-authored-by: Junwei Deng <[email protected]>
Co-authored-by: xiangyuT <[email protected]>
Co-authored-by: Chaselzxy <[email protected]>
Co-authored-by: Xinyi Zhang <[email protected]>
Co-authored-by: theaperdeng <[email protected]>
Co-authored-by: Le-Zheng <[email protected]>
Co-authored-by: fuckyouNoob <[email protected]>
Co-authored-by: Yishuo Wang <[email protected]>
Co-authored-by: Guoqiong Song <[email protected]>
Co-authored-by: binbin <[email protected]>
Co-authored-by: Hu, Zhaojie <[email protected]>
Co-authored-by: glorysdj <[email protected]>
Co-authored-by: David smurf <[email protected]>
Co-authored-by: root <root@jiao-test.rfprax1kp2xefkkygl2c02qm5g.cx.internal.cloudapp.net>
Co-authored-by: Yina Chen <[email protected]>
Co-authored-by: WangBin <[email protected]>
Co-authored-by: mycprotein <[email protected]>
liu-shaojun pushed a commit that referenced this pull request Mar 25, 2024
@smurf-1119 @TheaperDeng
Chronos: train tcn model on gpu and speed up inference on cpu (#5594)
c73fc9f 
* add nano_gpu

* simplify the code

* simplify the code

* modify the code

* add some updates

* add document

Co-authored-by: theaperdeng <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@TheaperDeng TheaperDeng TheaperDeng left review comments

Assignees
No one assigned
Labels
Chronos
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@smurf-1119 @TheaperDeng