All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog.
catalyst-tune
for Config API added #1411
catalyst-run
for Config API support added #1406
- Logger API naming #1405
- Catalyst architecture simplification.
- #1395, #1396, #1397, #1398, #1399, #1400, #1401, #1402, #1403.
- Additional tests for different hardware accelerators setups. Please check out the
tests/pipelines
folder for more information. BackwardCallback
andBackwardCallbackOrder
as an abstraction on top ofloss.backward
. Now you could easily log model gradients or transform them beforeOptimizerCallback
.CheckpointCallbackOrder
forICheckpointCallback
.
- Minimal python version moved to
3.7
, minimal pytorch version moved to1.4.0
. - Engines rewritten on top of Accelerate. First, we found these two abstractions very close to each other. Second, Accelerate provides additional user-friendly API and more stable API for "Nvidia APEX" and "Facebook Fairscale" - it does not support them.
- SelfSupervisedRunner moved to the
examples
folder from the Catalyst API. The only Runners API, that will be supported in the future:IRunner
,Runner
,ISupervisedRunner
,SupervisedRunner
due to their consistency. If you are interested in any other Runner API - feel free to write your ownCustomRunner
and useSelfSupervisedRunner
as an example. Runner.{global/stage}_{batch/loader/epoch}_metrics
renamed toRunner.{batch/loader/epoch}_metrics
CheckpointCallback
rewritten from scratch.- Catalyst registry moved to full-imports-paths only.
- Logger API changed to receive
IRunner
for alllog_*
methods. - Metric API:
topk_args
renamed totopk
- Contrib API: init imports from
catalyst.contrib
- removed, usefrom catalyst.contrib.{smth} import {smth}
. Could be change to full-imports-only in future versions for stability. - All quickstarts, minimal examples, notebooks and pipelines moved to new version.
- Codestyle moved to
89
right margin. Honestly speaking, it's much easier to maintain Catalyst with89
right margin on MBP'16.
ITrial
removed.- Stages support removed. While we embrace stages in deep learning experiments, current hardware accelerators are not prepared well for such setups. Additionally, ~95% of dl pipelines are single-stage. Multi-stage runner support is under review. For multi-stage support, please define a
CustomRunner
with rewritten API. - Config/Hydra API support removed. Config API is under review. For now, you could write your own Config API with hydra-slayer if needed.
catalyst-dl
scripts removed. Without Config API we don't need them anymore.Nvidia Apex
,Fairscale
,Albumentations
,Nifti
,Hydra
requiremets removed.OnnxCallback
,PruningCallback
,QuantizationCallback
,TracingCallback
removed from callbacks API. Theese callbacks are under review now.
If you have any questions on the Catalyst 22 edition updates, please join Catalyst slack for discussion.
- MNIST dataset for SSL banchmark (#1368)
- MoveiLens 20M dataset #1336
- logger property for logging customization (#1372)
- MacridVAE example (#1363)
- SSL benchmark results (#1374)
- Neptune example (#1377)
- multi-node support for engines (#1364)
- RL examples update to last version (#1370)
- DDPLoaderWrapper updated to new version (#1385)
num_classes
for classification metrics became optional (#1379)- colab ci/cd update to new verion
requests
requirements forcatalyst[cv]
added (#1371)- loader step counter (#1374)
- detection example data preprocessing (#1369)
- gradient clipping with fp16 runs (#1378)
- config API fix for DDP runs (#1383)
- checkpoint creation for fp16 engines (#1382)
- MultiVAE RecSys example (#1340)
- Returned
resume
support - resolved #1193 (#1349) - Smoothing dice loss to contrib (#1344)
profile
flag forrunner.train
(#1348)- MultiDAE RecSys example (#1356)
SETTINGS.log_batch_metrics
,SETTINGS.log_epoch_metrics
,SETTINGS.compute_per_class_metrics
for framework-wise Metric & Logger APIs specification (#1357)log_batch_metrics
andlog_epoch_metrics
options for all available Loggers (#1357)compute_per_class_metrics
option for all available multiclass/label metrics (#1357)- pytorch benchmark script and simplified MNIST (#1360)
- A few framework simplifications were made (#1346):
catalyst-contrib
scripts reduced tocollect-env
andproject-embeddings
onlycatalyst-dl
scripts recuded torun
andtune
onlytransforms.
prefix deprecated for Catalyst-based transformscatalyst.tools
moved tocatalyst.extras
- task-dependent extensions from
catalyst.data
moved tocatalyst.contrib.data
catalyst.data.transforms
moved tocatalyst.contrib.data.transforms
Normalize
,ToTensor
transforms renamed toNormalizeImage
,ImageToTensor
- metric learning extensions moved to
catalyst.contrib.data
catalyst.contrib
moved to code-as-a-documentation developmentcatalyst[cv]
andcatalyst[ml]
extensions moved to flatten architecture design; examples:catalyst.contrib.data.dataset_cv
,catalyst.contrib.data.dataset_ml
catalyst.contrib
moved to flatten architecture design; exampels:catalyst.contrib.data
,catalyst.contrib.datasets
,catalyst.contrib.layers
,catalyst.contrib.models
,catalyst.contrib.optimizers
,catalyst.contrib.schedulers
- internal functionality moved to
***._misc
modules catalyst.utils.mixup
moved tocatalyst.utils.torch
catalyst.utils.numpy
moved tocatalyst.contrib.utils.numpy
- default logging logic moved from "batch & epoch" to "epoch"-only to save computation time during logging; to respecify, please use:
SETTINGS.log_batch_metrics=True/False
oros.environ["CATALYST_LOG_BATCH_METRICS"]
SETTINGS.log_epoch_metrics=True/False
oros.environ["CATALYST_LOG_EPOCH_METRICS"]
- default metrics computation moved from "per-class & aggregations" to "aggregations"-only to save computation time during logging; to respecify, please use:
SETTINGS.compute_per_class_metrics=True/False
oros.environ["CATALYST_COMPUTE_PER_CLASS_METRICS"]
- no transformations required for MNIST contrib dataset (#1360
- A few framework simplifications were made (#1346):
catalyst.contrib.pandas
catalyst.contrib.parallel
catalyst.contrib.models.cv
- a few
catalyst.utils.misc
functions catalyst.extras
removed from the public documentation
- documentation search error (21.10 only) (#1346)
- docs examples (#1362)
- Self-Supervised benchmark: (#1365), (#1361)
- RSquareLoss (#1313)
- Self-Supervised example updates: (#1305), (#1322), (#1325), (#1335)
- Albert training example (#1326)
- YOLO-X (new) detection example and refactoring (#1324)
TopKMetric
abstraction (#1330)
- simlified readme (#1312)
- improved DDP tutorial (#1327)
CMCMetric
renamed from<prefix>cmc<suffix><k>
to<prefix>cmc<k><suffix>
(#1330)
- Zero seed error (#1329)
- updated codestyle issues (#1331)
- TopK metrics: (#1330), (#1334), (#1339)
--expdir
param forcatalyst-dl run
(#1338)- ControlFlowCallback for distributed setup (#1341)
- CometLogger support (#1283)
- CometLogger examples (#1287)
- XLA docs (#1288)
- Contarstive loss functions:
NTXentLoss
(#1278),SupervisedContrastiveLoss
(#1293) - Self supervised learning:
ISelfSupervisedRunner
,SelfSupervisedConfigRunner
,SelfSupervisedRunner
,SelfSupervisedDatasetWrapper
(#1278) - SimCLR example (#1278)
- Superivised Contrastive example (#1293)
- extra warnings for runner-callbacks interaction (#1295)
CategoricalRegressionLoss
andQuantileRegressionLoss
to thecontrib
(#1295)- R2 score metric (#1274)
- Improved
WandbLogger
to support artifacts and fix logging steps (#1309) - full
Runner
cleanup, with callbacks and loaders destruction, moved toPipelineParallelFairScaleEngine
only (#1295) HuberLoss
renamed toHuberLossV0
for the PyTorch compatibility (#1295)- codestyle update (#1298)
- BalanceBatchSampler - deprecated (#1303)
- RecSys loss functions:
AdaptiveHingeLoss
,BPRLoss
,HingeLoss
,LogisticLoss
,RocStarLoss
,WARPLoss
(#1269, #1282) - object detection examples (#1271)
- SklearnModelCallback (#1261)
- Barlow Twins example (#1261)
- TPU/XLA support (#1275)
- with updated example
- native
sync_bn
support for all available engines (#1275)- Torch, AMP, Apex, FairScale
- Registry moved to
hydra-slayer
(#1264)) - (#1275)
- batch metrics sync removed from ddp-runs to speedup training process
AccumulationMetric
renamed toAccumulativeMetric
- moved from
catalyst.metrics._metric
tocatalyst.metrics._accumulative
accululative_fields
renamed tokeys
- moved from
- PeriodicLoaderCallback docsting (#1279)
- matplotlib issue (#1272)
- sample counter for the loader (#1285)
- added
pre-commit
hook to run codestyle checker on commit (#1257) on publish
github action for docker and docs added (#1260)- MixupCallback and
utils.mixup_batch
(#1241) - Barlow twins loss (#1259)
- BatchBalanceClassSampler (#1262)
- make
expdir
incatalyst-dl run
optional (#1249) - Bump neptune-client from 0.9.5 to 0.9.8 in
requirements-neptune.txt
(#1251) - automatic merge for master (with Mergify) fixed (#1250)
- Evaluate loader custom model bug was fixed (#1254)
BatchPrefetchLoaderWrapper
issue with batch-based PyTorch samplers (#1262)- Adapted MlflowLogger for new config hierarchy (#1263)
- (#1230)
- FairScale support
- DeepSpeed support
utils.ddp_sync_run
function for synchronous ddp run- CIFAR10 and CIFAR100 datasets from torchvision (no cv-based requirements)
- Catalyst Engines demo
dataset_from_params
support in config API (#1231)- transform from params support for config API added (#1236)
- samplers from params support for config API added (#1240)
- recursive registry.get_from_params added (#1241)
- albumentations integration (#1238)
- Profiler callback (#1226)
- (#1230)
- loaders creation now wrapper with
utils.ddp_sync_run
forutils.ddp_sync_run
data preparation - runner support stage cleanup: loaders and callbacks will be deleted on the stage end
- Apex-based engines now support both APEXEngine and ApexEngine registry names
- loaders creation now wrapper with
- multiprocessing in minimal tests hotfix (#1232)
- Tracing callback hotfix (#1234)
- Engine hotfix for
predict_loader
(#1235) - (#1230)
- Hydra hotfix due to
1.1.0
version changes
- Hydra hotfix due to
HuberLoss
name conflict for pytorch 1.9 hotfix (#1239)
- Reinforcement learning tutorials (#1205)
- customization demo (#1207)
- FAQ docs: multiple input and output keys, engine tutorial (#1202)
- minimal Config API example (#1215)
- Distributed RL example (Catalyst.RL 2.0 concepts) (#1224)
- SklearnCallback as integration of sklearn metrics (#1198)
- tests moved to
tests
folder (#1208) - pipeline tests moved to
tests/pipelines
(#1215) - updated NeptuneLogger docstrings (#1223)
- customizing what happens in
train()
notebook (#1203) - transforms imports under catalyst.data (#1211)
- change layerwise to layerwise_params (#1210)
- add torch metrics support (#1195)
- add Config API support for BatchTransformCallback (#1209)
- Weights and Biases Logger (
WandbLogger
) (#1176) - Neptune Logger (
NeptuneLogger
) (#1196) log_artifact
method for logging arbitrary files like audio, video, or model weights toILogger
andIRunner
(#1196)
- Nifti Reader (NiftiReader) (#1151)
- CMC score and callback for ReID task (ReidCMCMetric and ReidCMCScoreCallback) (#1170)
- Market1501 metric learning datasets (Market1501MLDataset and Market1501QGDataset) (#1170)
- extra kwargs support for Engines (#1156)
- engines exception for unknown model type (#1174)
- a few docs to the supported loggers (#1174)
TensorboardLogger
switched fromglobal_batch_step
counter toglobal_sample_step
one (#1174)TensorboardLogger
logs loader metricon_loader_end
rather thanon_epoch_end
(#1174)prefix
renamed tometric_key
forMetricAggregationCallback
(#1174)micro
,macro
andweighted
aggregations renamed to_micro
,_macro
and_weighted
(#1174)BatchTransformCallback
updated (#1153)
- auto
torch.sigmoid
usage formetrics.AUCMetric
andmetrics.auc
(#1174)
- hitrate calculation issue (#1155)
- ILoader wrapper usage issue with Runner (#1174)
- counters for ddp case (#1174)
- minimal requirements issue (#1147)
- nested dicts in
loaders_params
/samplers_params
overriding (#1150)
- Additive Margin SoftMax(AMSoftmax) (#1125)
- Generalized Mean Pooling(GeM) (#1084)
- Key-value support for CriterionCallback (#1130)
- Engine configuration through cmd (#1134)
- Extra utils for thresholds (#1134)
- Added gradient clipping function to optimizer callback (1124)
- FactorizedLinear to contrib (1142)
- Extra init params for
ConsoleLogger
(1142) - Tracing, Quantization, Onnx, Pruninng Callbacks (1127)
- CriterionCallback now inherits from BatchMetricCallback #1130)
- united metrics computation logic
- Data-Model device sync and
Engine
logic duringrunner.predict_loader
(#1134) - BatchLimitLoaderWrapper logic for loaders with shuffle flag (#1136)
- config description in the examples (1142)
- Config API deprecated parsings logic (1142) (1138)
- RecSys metrics Top_k calculations (#1140)
_key_value
for schedulers in case of multiple optimizers (#1146)
[21.03] - 2021-03-13 (#1095)
Engine
abstraction to support various hardware backends and accelerators: CPU, GPU, multi GPU, distributed GPU, TPU, Apex, and AMP half-precision training.Logger
abstraction to support various monitoring tools: console, tensorboard, MLflow, etc.Trial
abstraction to support various hyperoptimization tools: Optuna, Ray, etc.Metric
abstraction to support various of machine learning metrics: classification, segmentation, RecSys and NLP.- Full support for Hydra API.
- Full DDP support for Python API.
- MLflow support for metrics logging.
- United API for model post-processing: tracing, quantization, pruning, onnx-exporting.
- United API for metrics: classification, segmentation, RecSys, and NLP with full DDP and micro/macro/weighted/etc aggregations support.
Experiment
abstraction merged intoRunner
one.- Runner, SupervisedRunner, ConfigRunner, HydraRunner architectures and dependencies redesigned.
- Internal settings and registry mechanisms refactored to be simpler, user-friendly and more extendable.
- Bunch of Config API test removed with Python API and pytest.
- Codestyle now supports up to 99 symbols per line :)
- All callbacks/runners moved for contrib to the library core if was possible.
Runner
abstraction simplified to store only current state of the experiment run: all validation logic was moved to the callbacks (by this way, you could easily select best model on various metrics simultaneously).Runner.input
andRunner.output
merged into unitedRunner.batch
storage for simplicity.- All metric moved from
catalyst.utils.metrics
tocatalyst.metrics
. - All metrics now works on scores/metric-defined-input rather that logits (!).
- Logging logic moved from
Callbacks
to appropriateLoggers
. KorniaCallbacks
refactored toBatchTransformCallback
.
- Lots of unnecessary contrib extensions.
- Transforms configuration support through Config API (could be returned in next releases).
- Integrated Python cmd command for model pruning, swa, etc (should be returned in next releases).
CallbackOrder.Validation
andCallbackOrder.Logging
- All 2020 year backward compatibility fixes and legacy support.
- Docs rendering simplified.
- LrFinderCallback.
Release docs, Python API minimal examples, Config/Hydra API example.
- Inference mode for face layers (#1045)
- Fix bug in
OptimizerCallback
when mixed-precision params set both: in callback arguments and in distributed_params (#1042)
- CVS Logger (#1005)
- DrawMasksCallback (#999)
- (#1002)
- a few docs
- (#998)
reciprocal_rank
metric- unified recsys metrics preprocessing
- (#1018)
- readme examples for all supported metrics under
catalyst.metrics
wrap_metric_fn_with_activation
for model outputs wrapping with activation- extra tests for metrics
- readme examples for all supported metrics under
- (#1039)
per_class=False
option for metrics callbacksPrecisionCallack
,RecallCallack
for multiclass problems- extra docs
- docs update (#1000)
AMPOptimizerCallback
andOptimizerCallback
were merged (#1007)- (#1017)
- fixed bug in
SchedulerCallback
- Log LRs and momentums for all param groups, not only for the first one
- fixed bug in
- (#1002)
tensorboard, ipython, matplotlib, pandas, scikit-learn
moved to optional requirementsPerplexityMetricCallback
moved tocatalyst.callbacks
fromcatalyst.contrib.callbacks
PerplexityMetricCallback
renamed toPerplexityCallback
catalyst.contrib.utils.confusion_matrix
renamed tocatalyst.contrib.utils.torch_extra
- many parts of
catalyst.data
moved tocatalyst.contrib.data
catalyst.data.scripts
moved tocatalyst.contrib.scripts
catalyst.utils
,catalyst.data.utils
andcatalyst.contrib.utils
restructuredReaderSpec
renamed toIReader
SupervisedExperiment
renamed toAutoCallbackExperiment
- gain functions renamed for
dcg
/ndcg
metrics (#998) - (#1014)
- requirements respecification:
catalyst[cv]
,catalyst[dev]
,catalyst[log]
,catalyst[ml]
,catalyst[nlp]
,catalyst[tune]
- settings respecification
- extra tests for settings
- contrib refactoring
- requirements respecification:
- iou and dice metrics moved to per-class computation (#1031)
- (#1002)
KNNMetricCallback
sklearn
mode forConfusionMatrixLogger
catalyst.data.utils
- unnecessary
catalyst.tools.meters
- todos for unnecessary docs
- (#1014)
- transformers-based contrib (too unstable)
- (#1018)
- ClasswiseIouCallback/ClasswiseJaccardCallback as deprecated on (should be refactored in future releases)
- prevented modifying config during the experiment and runner initialization (#1004)
- a few test for RecSys MAP computation (#1018)
- leave batch size the same for default distributed training (#1023)
- (#1032)
- Apex: now you can use apex for multiple models training
- Apex: DataParallel is allowed for opt_level other than "O1"
- DCG, nDCG metrics (#881)
- MAP calculations #968
- hitrate calculations [#975] (#975)
- extra functions for classification metrics (#966)
OneOf
andOneOfV2
batch transforms (#951)precision_recall_fbeta_support
metric (#971)- Pruning tutorial (#987)
- BatchPrefetchLoaderWrapper (#986)
- DynamicBalanceClassSampler (#954)
- update Catalyst version to
20.10.1
for tutorials (#967) - added link to dl-course (#967)
IRunner
-> simplifiedIRunner
(#984)- docs were restructured (#985)
set_global_seed
moved fromutils.seed
toutils.misc
(#986)
BatchTransformCallback
- addnn.Module
transforms support (#951)- moved to
contiguous
view for accuracy computation (#982) - fixed torch warning on
optimizer.py:140
(#979)
- MRR metrics calculation (#886)
- docs for MetricCallbacks (#947)
- SoftMax, CosFace, ArcFace layers to contrib (#939)
- ArcMargin layer to contrib (#957)
- AdaCos to contrib (#958)
- Manual SWA to utils (#945)
- fixed path to
CHANGELOG.md
file and add information about unit test toPULL_REQUEST_TEMPLATE.md
([#955])(#955) catalyst-dl tune
config specification - now optuna params are grouped understudy_params
(#947)IRunner._prepare_for_stage
logic moved toIStageBasedRunner.prepare_for_stage
(#947)- now we create components in the following order: datasets/loaders, model, criterion, optimizer, scheduler, callbacks
MnistMLDataset
andMnistQGDataset
data split logic - now targets of the datasets are disjoint (#949)- architecture redesign (#953)
- experiments, runners, callbacks grouped by primitives under
catalyst.experiments
/catalyst.runners
/catalyst.callbacks
respectively - settings and typing moved from
catalyst.tools.*
tocatalyst.*
- utils moved from
catalyst.*.utils
tocatalyst.utils
- experiments, runners, callbacks grouped by primitives under
- swa moved to
catalyst.utils
(#963)
AMPOptimizerCallback
- fix grad clip fn support (#948)- removed deprecated docs types (#947) (#952)
- docs for a few files (#952)
- extra backward compatibility fixes (#963)
- Runner registry support for Config API (#936)
catalyst-dl tune
command - Optuna with Config API integration for AutoML hyperparameters optimization (#937)OptunaPruningCallback
alias forOptunaCallback
(#937)- AdamP and SGDP to
catalyst.contrib.losses
(#942)
- Config API components preparation logic moved to
utils.prepare_config_api_components
(#936)
MovieLens dataset
loader (#903)force
andbert-level
keywords tocatalyst-data text2embedding
(#917)OptunaCallback
tocatalyst.contrib
(#915)DynamicQuantizationCallback
andcatalyst-dl quantize
script for fast quantization of your model (#890)- Multi-scheduler support for multi-optimizer case (#923)
- Native mixed-precision training support (#740)
OptiomizerCallback
- flaguse_fast_zero_grad
for faster (and hacky) version ofoptimizer.zero_grad()
(#927)IOptiomizerCallback
,ISchedulerCallback
,ICheckpointCallback
,ILoggerCallback
as core abstractions for Callbacks (#933)- flag
USE_AMP
for PyTorch AMP usage (#933)
- autoresume option for Config API (#907)
- a few issues with TF projector (#917)
- batch sampler speed issue (#921)
- add apex key-value optimizer support (#924)
- runtime warning for PyTorch 1.6 (920)
- Apex synbn usage (920)
- Catalyst dependency on system git (922)
CMCScoreCallback
(#880)- kornia augmentations
BatchTransformCallback
(#862) average_precision
andmean_average_precision
metrics (#883)MultiLabelAccuracyCallback
,AveragePrecisionCallback
andMeanAveragePrecisionCallback
callbacks (#883)- minimal examples for multiclass and multilabel classification (#883)
- experimental TPU support (#893)
- add
Imagenette
,Imagewoof
, andImagewang
datasets (#902) IMetricCallback
,IBatchMetricCallback
,ILoaderMetricCallback
,BatchMetricCallback
,LoaderMetricCallback
abstractions (#897)HardClusterSampler
inbatch sampler (#888)
- all registries merged to one
catalyst.registry
(#883) mean_average_precision
logic merged withaverage_precision
(#897)- all imports moved to absolute (#905)
catalyst.contrib.data
merged tocatalyst.data
(#905)- {breaking} Catalyst transform
ToTensor
was renamed toImageToTensor
(#905) TracerCallback
moved tocatalyst.dl
(#905)ControlFlowCallback
,PeriodicLoaderCallback
moved tocatalyst.core
(#905)
log
parameter toWandbLogger
(#836)- hparams experiment property (#839)
- add docs build on push to master branch (#844)
WrapperCallback
andControlFlowCallback
(#842)BatchOverfitCallback
(#869)overfit
flag for Config API (#869)InBatchSamplers
:AllTripletsSampler
andHardTripletsSampler
(#825)
- Renaming (#837)
SqueezeAndExcitation
->cSE
ChannelSqueezeAndSpatialExcitation
->sSE
ConcurrentSpatialAndChannelSqueezeAndChannelExcitation
->scSE
_MetricCallback
->IMetricCallback
dl.Experiment.process_loaders
->dl.Experiment._get_loaders
LRUpdater
become abstract class (#837)calculate_confusion_matrix_from_arrays
changed params order (#837)dl.Runner.predict_loader
uses_prepare_inner_state
and cleansexperiment
(#863)toml
to the dependencies (#872)
crc32c
dependency (#872)
workflows/deploy_push.yml
failed to push some refs (#864).dependabot/config.yml
contained invalid details (#781)LanguageModelingDataset
(#841)global_*
counters inRunner
(#858)- EarlyStoppingCallback considers first epoch as bad (#854)
- annoying numpy warning (#860)
PeriodicLoaderCallback
overwrites best state (#867)OneCycleLRWithWarmup
(#851)
- docs structure were updated during (#822)
utils.process_components
moved fromutils.distributed
toutils.components
(#822)catalyst.core.state.State
merged tocatalyst.core.runner._Runner
(#823) (backward compatibility included)catalyst.core.callback.Callback
now works directly withcatalyst.core.runner._Runner
state_kwargs
renamed tostage_kwargs
- Circle loss implementation (#802)
- BatchBalanceSampler for metric learning and classification (#806)
CheckpointCallback
: new argumentload_on_stage_start
which acceptsstr
andDict[str, str]
(#797)- LanguageModelingDataset to catalyst[nlp] (#808)
- Extra counters for batches, loaders and epochs (#809)
TracerCallback
(#789)
CheckpointCallback
: additional logic for argumentload_on_stage_end
- acceptsstr
andDict[str, str]
(#797)- counters names for batches, loaders and epochs (#809)
utils.trace_model
: changed logic -runner
argument was changed topredict_fn
(#789)- redesigned
contrib.data
andcontrib.datasets
(#820) catalyst.utils.meters
moved tocatalyst.tools
(#820)catalyst.contrib.utils.tools.tensorboard
moved tocatalyst.contrib.tools
(#820)
- Added new docs and minimal examples (#747)
- Added experiment to registry (#746)
- Added examples with extra metrics (#750)
- Added VAE example (#752)
- Added gradient tracking (#679
- Added dependabot (#771)
- Added new test for Config API (#768)
- Added Visdom logger (#769)
- Added new github actions and templates (#777)
- Added
save_n_best=0
support for CheckpointCallback (#784) - Added new contrib modules for CV (#793)
- Added new github actions CI (#791)
- Changed
Alchemy
dependency (fromalchemy-catalyst
toalchemy
) (#748) - Changed warnings logic (#719)
- Github actions CI was updated (#754)
- Changed default
num_epochs
to 1 forState
(#756) - Changed
state.batch_in
/state.batch_out
tostate.input
/state.output
(#763) - Moved
torchvision
dependency fromcatalyst
tocatalyst[cv]
(#738))
- Fixed docker dependencies ($753)
- Fixed
text2embeddding
script (#722) - Fixed
utils/sys
exception (#762) - Returned
detach
method (#766) - Fixed timer division by zero (#749)
- Fixed minimal torch version (#775)
- Fixed segmentation tutorial (#778)
- Fixed Dockerfile dependency (#780)