GitLab

Add early stopping logic to ncf keras when desired threshold is met. Also change the default batch size to match the tuned hyperparams

* Add benchmarks with the --cloning flag to Resnet and NFC.

* Renamed cloning to clone_model_in_keras_dist_strat. Dropped a few tests that aren't essential.

* Fixed up the formatting after re-naming the flag to a much longer  name.  Thanks, lint.
* Fixed the lint error in nfc_common.py

* Remove contrib imports, or move them inline

* Use exposed API for FixedLenFeature

* Replace tf.logging with absl logging

* Change GFile to v2 APIs

* replace tf.logging with absl loggin in movielens

* Fixing an import bug

* Change gfile to v2 APIs in code

* Swap to keras optimizer v2

* Bug fix for optimizer

* Change tf.log to tf.keras.backend.log

* Change the loss function to keras loss

* convert another loss to keras loss

* Resolve comments and fix lint

* Add a doc string

* Fix existing tests and add new tests for DS

* Added tests for multi-replica

* Fix lint

* resolve comments

* make estimator run in tf2.0

* use compat v1 loss

* fix lint issue

* tmp commit

* tmp commit

* first attempt (without eval)

* Bug fixes

* bug fixes

* training done

* Loss NAN, no eval

* Loss weight problem solved

* resolve the NAN loss problem

* Problem solved. Clean up needed

* Added a todo

* Remove debug prints

* Extract get_optimizer to ncf_common

* Move metrics computation back to neumf; use DS.scope api

* Extract DS.scope code to utils

* lint fixes

* Move obtaining DS above producer.start to avoid race condition

* move pt 1

* move pt 2

* Update the run script

* Wrap keras_model related code into functions

* Update the doc for softmax_logitfy and change the method name

* Resolve PR comments

* working version with: eager, DS, batch and no masks

* Remove git conflict indicator

* move reshape to neumf_model

* working version, not converge

* converged

* fix a test

* more lint fix

* more lint fix

* more lint fixes

* more lint fix

* Removed unused imports

* fix test

* dummy commit for kicking of checks

* fix lint issue

* dummy input to kick off checks

* dummy input to kick off checks

* add collective to dist strat

* addressed review comments

* add a doc string

Add bisection based producer for increased scalability, enable fully deterministic data production, and use the materialized and bisection producer to check each other (via expected output md5's)

2nd half of rough replacement pass

fix dataset map functions

reduce bias in sample selection

cache pandas work on a daily basis

cleanup and fix batch check for multi gpu

multi device fix

fix treatment of eval data padding

print data producer

replace epoch overlap with padding and masking

move type and shape info into the producer class and update run.sh with larger batch size hyperparams

remove xla for multi GPU

more cleanup

remove model runner altogether

bug fixes

address subtle pipeline hang and improve producer __repr__

fix crash

fix assert

use popen_helper to create pools

add StreamingFilesDataset and abstract data storage to a separate class

bug fix

fix wait bug and add manual stack trace print

more bug fixes and refactor valid point mask to work with TPU sharding

misc bug fixes and adjust dtypes

address crash from decoding bools

fix remaining dtypes and change record writer pattern since it does not append

fix synthetic data

use TPUStrategy instead of TPUEstimator

minor tweaks around moving to TPUStrategy

cleanup some old code

delint and simplify permutation generation

remove low level tf layer definition, use single table with slice for keras, and misc fixes

missed minor point on removing tf layer definition

fix several bugs from recombinging layer definitions

delint and add docstrings

Update ncf_test.py. Section for identical inputs and different outputs was removed.

update data test to run against the new producer class

I've noticed sometimes the async process's pool processes do not die when ncf_main.py ends and kills the async process. This commit fixes the issue.

* Merges TPU-TC optimizations into HEAD.

* Split a line that went over 80 from a tab.

* Remove trailing whitespace.

The option is --nouse_estimator

--ml_perf now just changes the model to make it MLPerf compliant. --output_ml_perf_compliance_logging adds the MLPerf compliance logs.

The error message was:

absl.flags._exceptions.IllegalFlagValueError: flag --ml_perf=None: ('Non-boolean argument to boolean flag', 'None')

* first pass at __getattr__ abuse logger

* first pass at adding tags to NCF

* minor formatting updates

* fix tag name

* convert metrics to python floats

* getting closer...

* direct mlperf logs to a file

* small tweaks and add stitching

* update tags

* fix tag and add a sudo call

* tweak format of run.sh

* delint

* use distribution strategies for evaluation

* address PR comments

* delint and fix test

* adjust flag validation for xla

* add prefix to distinguish log stitching

* fix index bug

* fix clear cache for root user

* dockerize cache drop

* TIL some regex magic

* intermediate commit

finish replacing spillover with resampled padding

intermediate commit

* resolve merge conflict

* intermediate commit

* further consolidate the data pipeline

* complete first pass at data pipeline refactor

* remove some leftover code

* fix test

* remove resampling, and move train padding logic into neumf.py

* small tweaks

* fix weight bug

* address PR comments

* fix dict zip. (Reed led me astray)

* delint

* make data test deterministic and delint

* Reed didn't lead me astray. I just can't read.

* more delinting

* even more delinting

* use resampling for last batch padding

* pad last batch with unique data

* Revert "pad last batch with unique data"

This reverts commit cbdf46efcd5c7907038a24105b88d38e7f1d6da2.

* move padded batch to the beginning

* delint

* fix step check for synthetic data

* move flagfile into the cache_dir

* remove duplicate code

* delint

* Add --use_synthetic_data option to NCF.

* Add comment to _SYNTHETIC_BATCHES_PER_EPOCH

* Fix test

* Hopefully fix lint issue

* improve default handling for eval_batch_size

* return eval_batch_size default to None

* fix syntax error

* move evaluation from numpy to tensorflow

fix syntax error

don't use sigmoid to convert logits. there is too much precision loss.

WIP: add logit metrics

continue refactor of NCF evaluation

fix syntax error

fix bugs in eval loss calculation

fix eval loss reweighting

remove numpy based metric calculations

fix logging hooks

fix sigmoid to softmax bug

fix comment

catch rare PIPE error and address some PR comments

* fix metric test and address PR comments

* delint and fix python2

* fix test and address PR comments

* extend eval to TPUs

* bug fixes and add seed

* more random corrections

* make cleanup more robust

* return cleanup fn

* delint and address PR comments.

* delint and fix tests

* delinting is never done

* add pipeline hashing

* delint

* Fix convergence issues for MLPerf.

Thank you to @robieta for helping me find these issues, and for providng an algorithm for the `get_hit_rate_and_ndcg_mlperf` function.

This change causes every forked process to set a new seed, so that forked processes do not generate the same set of random numbers. This improves evaluation hit rates.

Additionally, it adds a flag, --ml_perf, that makes further changes so that the evaluation hit rate can match the MLPerf reference implementation.

I ran 4 times with --ml_perf and 4 times without. Without --ml_perf, the highest hit rates achieved by each run were 0.6278, 0.6287, 0.6289, and 0.6241. With --ml_perf, the highest hit rates were 0.6353, 0.6356, 0.6367, and 0.6353.

* fix lint error

* Fix failing test

* Address @robieta's feedback

* Address more feedback

* intermediate commit

* ncf now working

* reorder pipeline

* allow batched decode for file backed dataset

* fix bug

* more tweaks

* parallize false negative generation

* shared pool hack

* workers ignore sigint

* intermediate commit

* simplify buffer backed dataset creation to fixed length record approach only. (more cleanup needed)

* more tweaks

* simplify pipeline

* fix misplaced cleanup() calls. (validation works\!)

* more tweaks

* sixify memoryview usage

* more sixification

* fix bug

* add future imports

* break up training input pipeline

* more pipeline tuning

* first pass at moving negative generation to async

* refactor async pipeline to use files instead of ipc

* refactor async pipeline

* move expansion and concatenation from reduce worker to generation workers

* abandon complete async due to interactions with the tensorflow threadpool

* cleanup

* remove performance_comparison.py

* experiment with rough generator + interleave pipeline

* yet more pipeline tuning

* update on-the-fly pipeline

* refactor preprocessing, and move train generation behind a GRPC server

* fix leftover call

* intermediate commit

* intermediate commit

* fix index error in data pipeline, and add logging to train data server

* make sharding more robust to imbalance

* correctly sample with replacement

* file buffers are no longer needed for this branch

* tweak sampling methods

* add README for data pipeline

* fix eval sampling, and vectorize eval metrics

* add spillover and static training batch sizes

* clean up cruft from earlier iterations

* rough delint

* delint 2 / n

* add type annotations

* update run script

* make run.sh a bit nicer

* change embedding initializer to match reference

* rough pass at pure estimator model_fn

* impose static shape hack (revisit later)

* refinements

* fix dir error in run.sh

* add documentation

* add more docs and fix an assert

* old data test is no longer valid. Keeping it around as reference for the new one

* rough draft of data pipeline validation script

* don't rely on shuffle default

* tweaks and documentation

* add separate eval batch size for performance

* initial commit

* terrible hacking

* mini hacks

* missed a bug

* messing about trying to get TPU running

* TFRecords based TPU attempt

* bug fixes

* don't log remotely

* more bug fixes

* TPU tweaks and bug fixes

* more tweaks

* more adjustments

* rework model definition

* tweak data pipeline

* refactor async TFRecords generation

* temp commit to run.sh

* update log behavior

* fix logging bug

* add check for subprocess start to avoid cryptic hangs

* unify deserialize and make it TPU compliant

* delint

* remove gRPC pipeline code

* fix logging bug

* delint and remove old test files

* add unit tests for NCF pipeline

* delint

* clean up run.sh, and add run_tpu.sh

* forgot the most important line

* fix run.sh bugs

* yet more bash debugging

* small tweak to add keras summaries to model_fn

* Clean up sixification issues

* address PR comments

* delinting is never over

* begin branch

* finish download script

* rename download to dataset

* intermediate commit

* intermediate commit

* misc tweaks

* intermediate commit

* intermediate commit

* intermediate commit

* delint and update census test.

* add movie tests

* delint

* fix py2 issue

* address PR comments

* intermediate commit

* intermediate commit

* intermediate commit

* finish wide deep transition to vanilla movielens

* delint

* intermediate commit

* intermediate commit

* intermediate commit

* intermediate commit

* fix import

* add default ncf csv construction

* change default on download_if_missing

* shard and vectorize example serialization

* fix import

* update ncf data unittests

* delint

* delint

* more delinting

* fix wide-deep movielens serialization

* address PR comments

* add file_io tests

* investigate wide-deep test failure

* remove hard coded path and properly use flags.

* address file_io test PR comments

* missed a hash_bucked_size

* Add DistributionStrategy to transformer model

* add num_gpu flag

* Calculate per device batch size for transformer

* remove reference to flags_core

* Add synthetic data option to transformer

* fix typo

* add import back in

* Use hierarchical copy

* address PR comments

* lint

* fix spaces

* group train op together to fix single GPU error

* Fix translate bug (sorted_keys is a dict, not a list)

* Change params to a default dict (translate.py was throwing errors because params didn't have the TPU parameters.)

* Address PR comments. Removed multi gpu flag + more

* fix lint

* fix more lints

* add todo for Synthetic dataset

* Update docs

* Add new test ID and test env info to the benchmark run.

* Fix test.

* Fix lint

* Address review comment.

* Update benchmark logger to update the run status.

This is important for streaming upload to bigquery so that the
dashboard can ignore the 'running' benchmark at the moment since
its not finished yet.

* Move the run status into a separate table.

Also update the run status in the benchmark uploader and
BigqueryBenchmarkLogger.

* Insert instead of update for the benchmark status for file logger.

* Address review comments.

Update the logger to have benchmark context, which will update the
run status accordingly.

* Fix broken tests.

* Move the benchmark logger context to main function.

* Fix tests.

* Update the rest of the models to use the context in main.

* Delint.

* Add unit test, official flags, and benchmark logs

* Fix checking errors

* Reorder imports to fix lints

* Address comments and correct model layers

* Add dataset checking

* Add recommendation model

* Fix pylints check error

* Rename file

* Address comments, update input pipeline, and add distribution strategy

* Fix import error

* Address more comments

* Fix lints