v1.2.10

What's Changed

• Remove geobenchv2 from test and vllm optional deps by @romeokienzler in #1208

Full Changelog: v1.2.9...v1.2.10

v1.2.9

What's Changed

• feat: Add DETR, Deformable DETR, and RF-DETR object detection models by @ahmedemam576 in #1165
• Add xView object-detection starter pipeline (preprocessing + data loa… by @ahmedemam576 in #1175
• feat(xview2): add TerraMind building-damage segmentation example for the XView2 dataset by @ahmedemam576 in #1188
• Replace private imports of TorchGeo models by @adamjstewart in #1196
• Skip tiled inference when input fits in a single chip by @blumenstiel in #1195
• Updated vLLM serving example by @christian-pinto in #1107
• update docs deploy workflow by @Isabelle-Wittmann in #1182
• fix(rfdetr): support 2/3-value backbone returns by @ahmedemam576 in #1205
• Added temporary fix to support vLLM >0.21.0 and <=0.24.0 by @christian-pinto in #1203
• Fix: Resolve KeyError and NameError in TorchGeo backbones weight loading by @clementgilli in #1200
• Support CPU training, add installation commands, download data and embeddings if not present by @romeokienzler in #1186
• integrationtests: make /dccstor data paths overridable via TERRATORCH_DATA_ROOT by @romeokienzler in #1206
• Bump version to v1.2.9 by @romeokienzler in #1207

New Contributors

• @clementgilli made their first contribution in #1200

Full Changelog: v1.2.8...v1.2.9

v1.2.8

What's Changed

• Add generic data tiling wrapper to solve padding consistency by @ahmedemam576 in #1140
• Update README links and badges to TorchGeo by @Isabelle-Wittmann in #1178
• update links to torchgeo by @Isabelle-Wittmann in #1179
• Fix denormalize by @blumenstiel in #1173
• Bump tj-actions/changed-files from 42 to 47 by @dependabot[bot] in #1118
• Fix typo in readme by @Isabelle-Wittmann in #1181
• Add Ahmed Emam to MAINTAINERS.md by @ahmedemam576 in #1185
• setuptools: adopt PEP 639 license metadata by @adamjstewart in #1184
• Move tacoreader to optional dependencies by @rosielickorish in #1189
• Add support to IOProcessor multi worker execution for Prithvi by @mgazz in #1191
• Bump version from 1.2.7 to 1.2.8 by @romeokienzler in #1193

Full Changelog: v1.2.7...v1.2.8

v1.2.5

What's Changed

• Custom modules set using TERRATORCH_CUSTOM_MODULE_PATH by @rosielickorish in #1133
• Extended support to vLLM>0.16.0 by @christian-pinto in #1136
• Refactor deployment step in publish-pypi.yml by @romeokienzler in #1137
• Add terracodec models by @Isabelle-Wittmann in #1142
• Update terratorch version to 1.2.4 by @Isabelle-Wittmann in #1144
• Improve integration tests execution script by @christian-pinto in #1141
• Fix plotting in multimodal datamodule by @blumenstiel in #1147
• update imports by @Isabelle-Wittmann in #1148
• update mps batchnorm fix by @Isabelle-Wittmann in #1149
• Fix typo in CONTRIBUTING.md "erraTorch" -> "TerraTorch" by @canmike in #1150
• Fix TiM chained generation by @blumenstiel in #1145
• Updated tox.ini to match changes in vLLM tests by @christian-pinto in #1153
• Bump version to v1.2.5 and update authors by @romeokienzler in #1154

Full Changelog: v1.2.4...v1.2.5

v1.2.4

What's Changed

• Add Geobench v2 predict loader by @Isabelle-Wittmann in #1097
• release v1.2.2 by @romeokienzler in #1103
• Embedding Georeferencing Updates by @Isabelle-Wittmann in #1102
• Vllm doc update 121 by @romeokienzler in #1110
• Vllm doc update 121 by @romeokienzler in #1115
• consolidate docs by @romeokienzler in #1116
• Delete .github/workflows/docs.yaml by @romeokienzler in #1117
• add torchgeo softcon by @Isabelle-Wittmann in #1114
• Embedding Updates for compatibility with NeuCo-Bench by @Isabelle-Wittmann in #1112
• add custom modules example by @romeokienzler in #1108
• Test/integration test fix v2 by @rosielickorish in #1120
• Improve integration test script by @rosielickorish in #1122
• Update pyproject.toml with new tag by @rosielickorish in #1124
• Fixed logic for tortilla file vs data roots by @thomas-mattsson in #1105
• Force update aliases during docs deployment by @romeokienzler in #1125
• Fix prediction plot by @blumenstiel in #1128
• Updates to vLLM plugins by @christian-pinto in #1127
• Changed version requirements for segmentation-models-pytorch by @christian-pinto in #1129
• Improve integration test script by @rosielickorish in #1130
• Bump version from v1.2.3pre1 to v1.2.4pre1 by @romeokienzler in #1131
• Update version to v1.2.4 in pyproject.toml by @romeokienzler in #1135

New Contributors

• @rosielickorish made their first contribution in #1120

Full Changelog: v1.2.2...v1.2.4

v1.2.1-hotfix

Full Changelog of hotfix: v1.2...v1.2.1

v1.2.2

What's Changed

• Update multimodal transform types and docs by @blumenstiel in #1046
• Updating instructions for installation as developer by @Joao-L-S-Almeida in #1050
• release v1.2, deprecate examples by @romeokienzler in #1047
• added NAIP standardization by @jhnnsjkbk1 in #1049
• Pin wandb by @blumenstiel in #1052
• Add migration script for TerraTorch YAML to v1.2 by @romeokienzler in #1054
• Embedding Workflows Integrations by @Isabelle-Wittmann in #1048
• Fix/basetask plotting by @Isabelle-Wittmann in #1051
• extend mlp decoder to 2d inputs by @Isabelle-Wittmann in #1060
• fixes to support vLLM >= 0.12 by @christian-pinto in #1067
• Add Isabelle Wittmann and pyarrow to dependencies by @romeokienzler in #1068
• Delete .github/workflows/test.yaml by @romeokienzler in #1071
• Enhance pull request template with testing instructions by @romeokienzler in #1072
• Add workflow to check pull request task lists by @romeokienzler in #1073
• Add new IO processor for Terramind Segmentation tasks by @christian-pinto in #1058
• Restoring ConvNext by @Joao-L-S-Almeida in #807
• Updated TM IO Processor Plugin to support the latest vLLM release by @christian-pinto in #1078
• Delete .github/workflows/prevent_unchecked_pr_tasks.yml by @romeokienzler in #1076
• Update examples by @Isabelle-Wittmann in #1075
• Fix biomassters dataset by @enomis-dev in #1036
• Add run_tests.sh script for automated testing by @romeokienzler in #1074
• [WIP] Builds on top of #969 — adds custom loss for classification tasks by @ltlinh16 in #986
• Building the package and testing it inside a container by @Joao-L-S-Almeida in #1016
• Fixed loss selection in task by @blumenstiel in #1028
• Revert "Fixed loss selection in task" by @romeokienzler in #1082
• Update embedding workflow readme & E2S ref by @Isabelle-Wittmann in #1081
• Update run_tests.sh by @romeokienzler in #1085
• Remove ignore index classification by @romeokienzler in #1084
• Fix multitemporal dataset by @Isabelle-Wittmann in #1083
• Fix test multitemporal dataset by @Isabelle-Wittmann in #1088
• add tests to require approval, only test on 3.13 by @romeokienzler in #1089
• Update revert padding to use center crop by @Isabelle-Wittmann in #1090
• Fix TM caption tokenizer by @blumenstiel in #1069
• Added vLLM prithvi serving example and fixed vLLM dependency by @christian-pinto in #1092
• Add/tests by @romeokienzler in #1012
• Refactor/unified normalize function by @ahmedemam576 in #1086
• docs: add notebook explaining 4 levels of model abstraction-how to use terra torch's models in a notebook by @ahmedemam576 in #1070
• add augmentation for object detection plus example by @romeokienzler in #1066
• add wxc unit test plus two fixes by @romeokienzler in #556
• [WIP] remove sync dist by @blumenstiel in #1004
• Support of Terrakit tortillas for segmentation tasks by @thomas-mattsson in #1059
• Docs: add AED object detection explanation by @ahmedemam576 in #1063
• Add script to run LSF integration tests with logging by @romeokienzler in #1055
• Delete examples/notebooks/od_data_augmentation.ipynb by @romeokienzler in #1100
• Various fixes to vLLM plugins by @christian-pinto in #1101

New Contributors

• @jhnnsjkbk1 made their first contribution in #1049
• @enomis-dev made their first contribution in #1036
• @ltlinh16 made their first contribution in #986
• @ahmedemam576 made their first contribution in #1086
• @thomas-mattsson made their first contribution in #1059

Full Changelog: v1.2.1...v1.2.2

v1.2

What's Changed

• Change YouTube link and update video description by @romeokienzler in #1008

• Add diffusers dependency with version constraint by @romeokienzler in #1010

• Terramind integration test by @WanjiruCate in #1013

• Geobench2-visualization by @naomi-simumba in #1021

• update type hint in generic multimodal datamodule by @Isabelle-Wittmann in #1025

• It will be moved to the documentation by @Joao-L-S-Almeida in #1006

• ScalarRegressionTask by @leonardijulia in #929

• Update TerraMind dtypes by @blumenstiel in #1033

• Improvments for multimodal data by @blumenstiel in #1009

• Cleaning up cache in two stages to avoid OSError by @Joao-L-S-Almeida in #1041

• Deprecated classification head warning by @blumenstiel in #1039

• Always generated float16 dummy inputs (vLLM default) by @christian-pinto in #1044

• Updating Discord link to avoid expiration by @Joao-L-S-Almeida in #1045

• fixed issues with the tutorial. missing model and broken file import by @Mdmdma in #1042

• Update/albucore-jsonargparse-pytorch by @romeokienzler in #1040

• Improvements to PixelWiseModel Padding Handling

This update fixes issues in how padding is handled by PixelWiseModel's forward method, by moving automatic padding from bottom right to all corners, with the important caveat that this change is not backward compatible with TerraTorch versions prior to 1.2. This means that models must be re-trained or fine-tuned with TerraTorch ≥ 1.2 for inference, if the model was trained using padding (training image size not divisible by 2 * patch size, i.e. 32 pixels for most models).

For more technical details, see Issue #1079

New Contributors

• @leonardijulia made their first contribution in #929
• @Mdmdma made their first contribution in #1042

Full Changelog: v1.1.1...v1.2

v1.1.1

What's Changed

• Update Slack invite link in README by @romeokienzler in #891
• Remove pytest and pytest-sugar from pyproject.toml by @romeokienzler in #893
• Remove unnecessary dtype definitions by @mgazz in #892
• Add finetuning code for terramind tiny and the AED Elephant Dataset by @romeokienzler in #847
• add tm.tiny demo by @romeokienzler in #896
• add prithvi.tiny webapp demo by @romeokienzler in #899
• Fixing tests by @Joao-L-S-Almeida in #898
• adds config for prithvi tiny finetune on m-chesapeake by @PedroConrado in #900
• fixes examples/confs/prithvi_tiny_finetune_chesapeake.yaml name by @PedroConrado in #901
• Fix GitHub license detection by @adamjstewart in #902
• Restoring downloads counter by @Joao-L-S-Almeida in #894
• Update test workflow to ignore changes in the integration tests by @Joao-L-S-Almeida in #886
• Updates to web app examples by @Zoe-Gathercole1 in #904
• Updated vLLM Segmentation IO Processor plugin and added tests by @christian-pinto in #895
• change import to support new package name of tt-iterate by @leotizzei in #918
• Adding badge for conda by @Joao-L-S-Almeida in #919
• Special Exception to handle models missing from the registry by @Joao-L-S-Almeida in #921
• Trying to fix number format issue by @Joao-L-S-Almeida in #916
• geobench v2 wrapper by @naomi-simumba in #710
• add elephant object detection fine tuning by @romeokienzler in #848
• Fix condition initialization tokenizer file by @Evameijling in #926
• fix/dependencies by @blumenstiel in #931
• Update vLLM tiled parameters by @blumenstiel in #930
• Fixing initialization of the tiling parameters in vLLM plugins by @christian-pinto in #933
• Add boundary-loss by @blumenstiel in #897
• Update TerraMind model param warnings by @blumenstiel in #911
• Enhance CI workflow with submodule updates and coverage checks by @romeokienzler in #932
• Updating README by @Joao-L-S-Almeida in #937
• Add diff_subset option temporal wrapper, update doc and tutorial notebook by @Isabelle-Wittmann in #957
• Update main code in docs by @blumenstiel in #959
• Moving coverage analysis to external github runner by @romeokienzler in #958
• Add TerraMind embedding generation example notebook - fix by @GieziJo in #837
• Badge was removed by @Joao-L-S-Almeida in #962
• Mention Iterate tool in README by @romeokienzler in #963
• Updating README by @Joao-L-S-Almeida in #965
• Fix indent error, integrate commented out tests, and improve docs by @Isabelle-Wittmann in #956
• Update README.md by @biancazadrozny in #968
• Updating README by @Joao-L-S-Almeida in #966
• Remove xformers from TerraMind tokenizers by @blumenstiel in #971
• Geobench v2 od by @paolofraccaro in #923
• Updating the github.io address by @Joao-L-S-Almeida in #984
• Updating README by @Joao-L-S-Almeida in #985
• Async plugin update by @christian-pinto in #952
• Geobench must be installed by @Joao-L-S-Almeida in #980
• Updating domain by @Joao-L-S-Almeida in #988
• move from slack to discord by @romeokienzler in #983
• Set pretrained false in tests by @blumenstiel in #991
• Separating vllm dependencies from tests dependencies by @christian-pinto in #996
• Geobench from GH by @Joao-L-S-Almeida in #989
• Periodical synchronization between docs and main by @Joao-L-S-Almeida in #987
• Updating segmentation-pytorch-models by @Joao-L-S-Almeida in #995
• Update README.md by @romeokienzler in #999
• Update geobenchv2 dependency in pyproject.toml by @romeokienzler in #997
• update links in readme by @romeokienzler in #1002
• Enforcing installation of geobench for the tests by @Joao-L-S-Almeida in #1001
• Configuration for tox by @Joao-L-S-Almeida in #994
• add inria preprocessing for OD by @romeokienzler in #981
• Fix Geobench2, compute-statistics, multimodal transform by @blumenstiel in #998
• Mute albumentations by @blumenstiel in #1005
• Support for inference using Surya downstream models by @Joao-L-S-Almeida in #967
• Running the tests inside container environments (for local execution) by @Joao-L-S-Almeida in #924
• Update pyproject.toml by @romeokienzler in #1007

New Contributors

• @adamjstewart made their first contribution in #902
• @Zoe-Gathercole1 made their first contribution in #904
• @Evameijling made their first contribution in #926

Full Changelog: v1.1...v1.1.1

v1.1

What's Changed

• Update DOFA Backbone for Arbitrary Image Sizes by @xiong-zhitong in #577
• restore pyproject.toml by @romeokienzler in #689
• Add add object detection to TerraTorch around torchvision-detection by @paolofraccaro in #660
• add reference to arXiv by @romeokienzler in #691
• Testing for python 3.13 by @Joao-L-S-Almeida in #560
• Added support for temporal segmentation data in encoder decoder factory by @rijuld in #355
• Add TensorRT support for performance benchmarking by @romeokienzler in #695
• Fix batch size in tasks by @blumenstiel in #698
• Update tiled inference by @blumenstiel in #699
• Update default metrics by @blumenstiel in #697
• Periodical documentation synchronization with main by @Joao-L-S-Almeida in #690
• Experimental stage for the ruff linting/reformatting by @Joao-L-S-Almeida in #684
• Basic method to explore model cards by @Joao-L-S-Almeida in #687
• add era5 test by @romeokienzler in #700
• Update email in toml file by @blumenstiel in #707
• Remove MultiMAE by @blumenstiel in #708
• fix indices by @naomi-simumba in #717
• Update generic multimodal dataset by @blumenstiel in #705
• unpack sequential scheduler by @naomi-simumba in #656
• add missing import: from torchmetrics.wrappers import ClasswiseWrapper by @leotizzei in #713
• add wandb support for integration tests by @romeokienzler in #726
• Fix/temporal wrapper by @paolofraccaro in #725
• Non-divisible tokens by @Joao-L-S-Almeida in #720
• Update class-wise metric names by @blumenstiel in #733
• Update README.md by @KonstiDE in #746
• Remove img_grep & label_grep update from config_deploy.yaml used by Studio by @WanjiruCate in #745
• Fix merge method dict for tm by @blumenstiel in #744
• Clay v1.5 preview by @romeokienzler in #738
• Add input dict support to tiled inference by @blumenstiel in #751
• add funding credits by @romeokienzler in #752
• Squeezing ground truth with innapropriate formats by @Joao-L-S-Almeida in #753
• Dtypes and loss by @Joao-L-S-Almeida in #749
• Fix/custom modules by @Joao-L-S-Almeida in #742
• The output of prediction can be None. by @Joao-L-S-Almeida in #747
• Add example for exporting models to TorchJIT by @romeokienzler in #757
• Onboarding Galileo by @Joao-L-S-Almeida in #659
• Update slack invite link by @romeokienzler in #758
• Fixing dependencies installation during tests by @Joao-L-S-Almeida in #766
• Increasing timeout by @Joao-L-S-Almeida in #767
• Fix/760 by @Joao-L-S-Almeida in #764
• Fix/761 by @Joao-L-S-Almeida in #765
• Label grep still needs be informed in the prediction mode by @Joao-L-S-Almeida in #769
• Update/release scripts by @Joao-L-S-Almeida in #776
• tiled inference on validation by @naomi-simumba in #729
• Updates for TerraMind by @blumenstiel in #676
• Trying to avoid the device be returned as negative by @Joao-L-S-Almeida in #779
• It's necessary to overwrite the find_classes method the generic scalar dataset by @Joao-L-S-Almeida in #770
• Testing workflow to deploy documentation by @Joao-L-S-Almeida in #790
• Fix metrics by @blumenstiel in #789
• Update integration tests by @WanjiruCate in #785
• Not the correct branch by @Joao-L-S-Almeida in #792
• Fix syntax by @Joao-L-S-Almeida in #795
• Fix/download stats by @Joao-L-S-Almeida in #787
• pretraine=False by @Joao-L-S-Almeida in #798
• Missing dependencies for deploying documentation by @Joao-L-S-Almeida in #800
• Fix docs workflow by @blumenstiel in #801
• Add test for multimodal datamodule by @blumenstiel in #783
• Configuring environment to push to github.io by @Joao-L-S-Almeida in #802
• Add an inference runner to build a general model executor in vLLM by @mgazz in #806
• Add secret to GH Actions bot by @Joao-L-S-Almeida in #804
• Fix image_grep checks in multimodal data module by @blumenstiel in #809
• Adjust to try fixing broken docs by @Joao-L-S-Almeida in #813
• Update docs by @Joao-L-S-Almeida in #794
• Making Galileo optional by @Joao-L-S-Almeida in #812
• Reduce tests by @blumenstiel in #819
• Update docs branch by @blumenstiel in #821
• Update Slack invite link in README by @romeokienzler in #824
• Trying to update albumentations by @Joao-L-S-Almeida in #803
• Fix config path in test_models_fit function by @romeokienzler in #828
• Add disclaimer about model usage and responsibilities by @romeokienzler in #827
• Fix ReshapeTokensToImage by @blumenstiel in #825
• Feature/embedding-generation by @Isabelle-Wittmann in #810
• add TerraMind embedding generation example notebook by @GieziJo in #833
• Remove initialization logic by @romeokienzler in #844
• improve verbosity on tests by @romeokienzler in #840
• Add Integration test for surya by @WanjiruCate in #852
• Updating dependabot rules by @Joao-L-S-Almeida in #835
• Adding support for Surya FM by @Joao-L-S-Almeida in #820
• Missing diffusers by @Joao-L-S-Almeida in #818
• Improve/data root handling by @Isabelle-Wittmann in #858
• Adding TM tokenizers as backbones by @blumenstiel in #856
• Bump actions/checkout from 3 to 5 by @dependabot[bot] in #808
• Add/temporalwrapper doc by @Isabelle-Wittmann in #843
• Removing any reference to Galileo by @Joao-L-S-Almeida in #861
• Add img_suffix to fix failing integration tests by @WanjiruCate in #862
• Fix tokenizer imports and make it optional by @blumenstiel in #863
• Pinning Surya tag by @Joao-L-S-Almeida in #864
• Trying to enforce the latest version of torchgeo by @Joao-L-S-Almeida in #860
• Integrate vLLM IO Processor plugins with terratorch + Segmentation task plugin by @christian-pinto in #867
• Add config generation script for running Terratorch models with vLLM by @mgazz in #868
• Fixes on top of #867 by @christian-pinto in #872
• Configure Mend for GitHub Enterprise by @ibm-mend-app[bot] in #870
• Make Sunya import optional by @blumenstiel in #877
• Configure integration tests by @WanjiruCate in #874
• Fixed NecksSequential by overloading forward for nn.Sequential by @christian-pinto in #878
• Update dtypes in terramind by @blumenstiel in #875
• Updating assertion by @Joao-L-S-Almeida in #882
• Update timm import by @blumenstiel in #883
• Update tm paths by @blumenstiel in #884
• Fixes for integr...