HEAD | 02a55a3 | [staticdocs][mmf] automatic update for docusaurus-plugin-internaldocs-fb@1.7.0 Differential Revision: D42920157 fbshipit-source-id: f4958274017b9e880c38ebbd6d7ea9411546e08d | 02 February 2023, 10:46:34 UTC |
refs/heads/0.1 | 5ad666f | Merge pull request #39 from HuaizhengZhang/patch-1 the extract command has a mistake | 08 March 2019, 04:39:38 UTC |
refs/heads/additional_validate | 2eb1d88 | [feat] PL last validation after training | 10 May 2021, 03:54:24 UTC |
refs/heads/airstore | 7ea5f8e | fix lint | 11 June 2021, 20:23:09 UTC |
refs/heads/audio_video_encoders | 96ecb11 | [feat] Adds audio (resnet18) and video (r2plus1d18) encoders (#879) Summary: This PR adds support for audio and video modality encoders to MMF. These can be used in conjunction with MMFTransformer. An example config has been added to showcase the usage. Pull Request resolved: https://github.com/facebookresearch/mmf/pull/879 Test Plan: Unit tests have been added. Reviewed By: ytsheng Differential Revision: D27804875 Pulled By: apsdehal fbshipit-source-id: 9f276dab2dc711fb8e5868a029f73c16083c1782 | 29 April 2021, 08:16:00 UTC |
refs/heads/config_mmft_doc | ff5c63f | [docs] Docs for various MMF Transformer configurations | 17 December 2020, 16:21:30 UTC |
refs/heads/debug_ln_masked_lm | 3ddd2aa | [feat] masked multi sentence bert_tokenizer processor | 19 October 2020, 19:21:56 UTC |
refs/heads/dependabot/npm_and_yarn/website/ansi-regex-4.1.1 | f805f3a | Bump ansi-regex from 4.1.0 to 4.1.1 in /website Bumps [ansi-regex](https://github.com/chalk/ansi-regex) from 4.1.0 to 4.1.1. - [Release notes](https://github.com/chalk/ansi-regex/releases) - [Commits](https://github.com/chalk/ansi-regex/compare/v4.1.0...v4.1.1) --- updated-dependencies: - dependency-name: ansi-regex dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com> | 02 February 2023, 10:49:39 UTC |
refs/heads/dependabot/npm_and_yarn/website/http-cache-semantics-4.1.1 | 3cf2f16 | Bump http-cache-semantics from 4.1.0 to 4.1.1 in /website Bumps [http-cache-semantics](https://github.com/kornelski/http-cache-semantics) from 4.1.0 to 4.1.1. - [Release notes](https://github.com/kornelski/http-cache-semantics/releases) - [Commits](https://github.com/kornelski/http-cache-semantics/commits) --- updated-dependencies: - dependency-name: http-cache-semantics dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com> | 02 February 2023, 10:49:47 UTC |
refs/heads/dependabot/npm_and_yarn/website/terser-5.16.2 | f6334f0 | Bump terser from 5.10.0 to 5.16.2 in /website Bumps [terser](https://github.com/terser/terser) from 5.10.0 to 5.16.2. - [Release notes](https://github.com/terser/terser/releases) - [Changelog](https://github.com/terser/terser/blob/master/CHANGELOG.md) - [Commits](https://github.com/terser/terser/compare/v5.10.0...v5.16.2) --- updated-dependencies: - dependency-name: terser dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com> | 02 February 2023, 10:49:40 UTC |
refs/heads/dependabot/npm_and_yarn/website/webpack-5.76.1 | 58c1862 | Bump webpack from 5.75.0 to 5.76.1 in /website Bumps [webpack](https://github.com/webpack/webpack) from 5.75.0 to 5.76.1. - [Release notes](https://github.com/webpack/webpack/releases) - [Commits](https://github.com/webpack/webpack/compare/v5.75.0...v5.76.1) --- updated-dependencies: - dependency-name: webpack dependency-type: indirect ... Signed-off-by: dependabot[bot] <support@github.com> | 15 March 2023, 10:29:03 UTC |
refs/heads/es | 79600ba | [feat][PL][4/N] best ckpt ghstack-source-id: fb6f40e23d5c5951aa58951b05601884922e10dc Pull Request resolved: https://github.com/facebookresearch/mmf/pull/996 | 30 June 2021, 18:34:45 UTC |
refs/heads/fix | 50d7ba4 | [fix] refactor regression | 21 June 2021, 23:45:04 UTC |
refs/heads/fix_base_model | bb2014b | [fix] minor change in base model ghstack-source-id: fb6f40e23d5c5951aa58951b05601884922e10dc Pull Request resolved: https://github.com/facebookresearch/mmf/pull/996 | 30 June 2021, 18:34:45 UTC |
refs/heads/fix_ci | f4b7566 | try skipping tests | 19 December 2022, 23:38:34 UTC |
refs/heads/fix_circleci_tests | c7cac1b | [fix] CircleCI GPU tests; sharded DDP tests - Due to various reasons (num_workers=1) etc sharded ddp's process destruction didn't work properly and gets stuck in between. This PR aims to fix that. | 29 April 2021, 23:00:03 UTC |
refs/heads/fix_lint | 7a67d7b | [fix,chores] Lint and wandb - Fixes lint issues introduced by recent PR - Fixes wandb issue as reported and fixed #1088 as author is unresponsive. Test Plan: CI | 14 October 2021, 19:31:24 UTC |
refs/heads/fixblack | 93d34e9 | [fix] black | 06 July 2021, 21:33:12 UTC |
refs/heads/fsdp | 2bfba58 | [WIP][feat] FSDP | 08 June 2021, 15:04:37 UTC |
refs/heads/fsdp_asg | 4347a0d | fsdp | 17 September 2021, 01:47:19 UTC |
refs/heads/fsdp_fairseq | c525893 | [fix] Fairseq support | 14 October 2021, 06:18:45 UTC |
refs/heads/fsdp_support | 63c18a1 | [feat] FSDP support for MMF | 29 September 2021, 18:58:24 UTC |
refs/heads/gh-pages | 19c460f | deploy: d1df04c4f4fc43afd52becd6d2ece601de6bc279 | 24 November 2021, 05:25:48 UTC |
refs/heads/gh/ebsmothers/1/base | dabf95f | Revert D30704069: [feat] Add a refiner head that can be used with MMFT Differential Revision: D30704069 (https://github.com/facebookresearch/mmf/commit/efa98757ccbecc4eb6a17532a38209679b7eb60b) Original commit changeset: 94e31778bd7b fbshipit-source-id: e6c6ffe345861b8bf05556c5d8f87613fc1263ba | 26 October 2021, 19:18:43 UTC |
refs/heads/gh/ryan-qiyu-jiang/30/base | da297e4 | Update base for Update on "[docs] Add ViLT model checkpoint instructions" Add ViLT model checkpoints to model zoo. Add ViLT finetuning and pretraining instructions to project docs. Differential Revision: [D32698147](https://our.internmc.facebook.com/intern/diff/D32698147) [ghstack-poisoned] | 13 December 2021, 21:11:52 UTC |
refs/heads/gh/ryan-qiyu-jiang/30/head | da07f13 | Update on "[docs] Add ViLT model checkpoint instructions" Add ViLT model checkpoints to model zoo. Add ViLT finetuning and pretraining instructions to project docs. Differential Revision: [D32698147](https://our.internmc.facebook.com/intern/diff/D32698147) [ghstack-poisoned] | 13 December 2021, 21:11:52 UTC |
refs/heads/gh/ryan-qiyu-jiang/30/orig | d412c25 | [docs] Add ViLT model checkpoint instructions Add ViLT model checkpoints to model zoo. Add ViLT finetuning and pretraining instructions to project docs. ghstack-source-id: 99480054d06835e7234da469564b63bd5a8ec6e7 Pull Request resolved: https://github.com/facebookresearch/mmf/pull/1161 | 13 December 2021, 21:11:52 UTC |
refs/heads/gh/ryan-qiyu-jiang/32/base | e583f3b | [feat] Add pytorchvideo encoder wrapper Add an encoder class that constructs any pytorchvideo model from config, and uses this model for its forward pass. Can load pretrained or random init models, based on config. [ghstack-poisoned] | 23 November 2021, 22:56:03 UTC |
refs/heads/gh/ryan-qiyu-jiang/32/head | fc0f8c0 | [docs] Add pytorchvideo docs Add pytorchvideo tutorial docs for using a pytorchvideo model as an encoder with TorchVideoEncoder class. [ghstack-poisoned] | 01 December 2021, 21:23:38 UTC |
refs/heads/gh/ryan-qiyu-jiang/32/orig | 74ad9d0 | [docs] Add pytorchvideo docs Add pytorchvideo tutorial docs for using a pytorchvideo model as an encoder with TorchVideoEncoder class. ghstack-source-id: 9cc594fc33357f86b45efe5db14e04f50f71804c Pull Request resolved: https://github.com/facebookresearch/mmf/pull/1163 | 01 December 2021, 21:23:41 UTC |
refs/heads/gh/ryan-qiyu-jiang/39/base | e8e2969 | Update base for Update on "[fix] Update UNITER VILLA checksums" Update UNITER and VILLA checksums for new checkpoints that include empty configs. Differential Revision: [D33064275](https://our.internmc.facebook.com/intern/diff/D33064275) [ghstack-poisoned] | 16 December 2021, 23:18:48 UTC |
refs/heads/gh/ryan-qiyu-jiang/39/head | b47d890 | Update on "[fix] Update UNITER VILLA checksums" Update UNITER and VILLA checksums for new checkpoints that include empty configs. Differential Revision: [D33064275](https://our.internmc.facebook.com/intern/diff/D33064275) [ghstack-poisoned] | 16 December 2021, 23:18:48 UTC |
refs/heads/gh/ryan-qiyu-jiang/39/orig | 3443b70 | [fix] Update UNITER VILLA checksums Update UNITER and VILLA checksums for new checkpoints that include empty configs. ghstack-source-id: 6876a49a051b8f82e6d4ec324afcb3802c907492 Pull Request resolved: https://github.com/facebookresearch/mmf/pull/1175 | 16 December 2021, 23:18:48 UTC |
refs/heads/gh/ryan-qiyu-jiang/40/base | 5cb8004 | Update base for Update on "[docs] Update UNITER project doc" Update doc to include VILLA citation and feature discrepancy explaination. Differential Revision: [D33064276](https://our.internmc.facebook.com/intern/diff/D33064276) [ghstack-poisoned] | 16 December 2021, 23:18:48 UTC |
refs/heads/gh/ryan-qiyu-jiang/40/head | 1ea94d2 | Update on "[docs] Update UNITER project doc" Update doc to include VILLA citation and feature discrepancy explaination. Differential Revision: [D33064276](https://our.internmc.facebook.com/intern/diff/D33064276) [ghstack-poisoned] | 16 December 2021, 23:18:48 UTC |
refs/heads/gh/ryan-qiyu-jiang/40/orig | e668551 | [docs] Update UNITER project doc Update doc to include VILLA citation and feature discrepancy explaination. ghstack-source-id: f696a850ac7a9c05b7860a4c4e86d531ed553138 Pull Request resolved: https://github.com/facebookresearch/mmf/pull/1176 | 16 December 2021, 23:18:48 UTC |
refs/heads/gh/ryan-qiyu-jiang/42/base | da07f13 | Update on "[docs] Add ViLT model checkpoint instructions" Add ViLT model checkpoints to model zoo. Add ViLT finetuning and pretraining instructions to project docs. Differential Revision: [D32698147](https://our.internmc.facebook.com/intern/diff/D32698147) [ghstack-poisoned] | 13 December 2021, 21:11:52 UTC |
refs/heads/gh/ryan-qiyu-jiang/42/head | e8991a8 | [docs] Add ViLT model code usage example Add an example of running ViLT from code outside of MMF cli at the end of the ViLT tutorial. Example is of ViLT vqa on raw image and text. [ghstack-poisoned] | 16 December 2021, 20:54:07 UTC |
refs/heads/gh/ryan-qiyu-jiang/42/orig | 3205392 | [docs] Add ViLT model code usage example Add an example of running ViLT from code outside of MMF cli at the end of the ViLT tutorial. Example is of ViLT vqa on raw image and text. ghstack-source-id: 4f7804ebe9ef9088197cbb83211e815219ecdfd8 Pull Request resolved: https://github.com/facebookresearch/mmf/pull/1179 | 16 December 2021, 20:54:11 UTC |
refs/heads/gh/ytsheng/2/base | b771f97 | [fix] black and isort (#991) Summary: Pull Request resolved: https://github.com/facebookresearch/mmf/pull/991 Reviewed By: vedanuj Differential Revision: D29448673 Pulled By: ytsheng fbshipit-source-id: f60d39fee67e331e09314e4f644b03dde1d50caf | 29 June 2021, 21:13:48 UTC |
refs/heads/gh/ytsheng/2/orig | 531a1f2 | [fix]: clean up lightning test code ghstack-source-id: a131a09f627773269e82ee63f790ee9cff8be3ef Pull Request resolved: https://github.com/facebookresearch/mmf/pull/995 | 30 June 2021, 18:34:41 UTC |
refs/heads/gh/ytsheng/21/base | 75ea3ab | [fix] Add zoo requirements to coco masked config (#1072) Summary: Add same zoo_requirements as coco dataset to masked_coco. Prevents file not found error while loading datasets. Pull Request resolved: https://github.com/facebookresearch/mmf/pull/1072 Reviewed By: apsdehal Differential Revision: D30727857 Pulled By: Ryan-Qiyu-Jiang fbshipit-source-id: b96b26063156d2bb463fa6d19e254a9f703f9242 | 02 September 2021, 19:29:28 UTC |
refs/heads/gh/ytsheng/21/orig | d656e49 | [fix] demjson depends on Setuptools 58.0.0, which has removed support for 2to3, so import would fail ghstack-source-id: cfc238e26e16cf62d5efdbd8a3e83860f40fd39f Pull Request resolved: https://github.com/facebookresearch/mmf/pull/1076 | 08 September 2021, 16:46:54 UTC |
refs/heads/gh/ytsheng/3/base | 2192818 | [fix]: clean up lightning test code [ghstack-poisoned] | 30 June 2021, 18:34:36 UTC |
refs/heads/gh/ytsheng/3/orig | f270b93 | [feat] PL train checkpoint ghstack-source-id: fb6f40e23d5c5951aa58951b05601884922e10dc Pull Request resolved: https://github.com/facebookresearch/mmf/pull/996 | 30 June 2021, 18:34:45 UTC |
refs/heads/gh/ytsheng/4/base | 273a56b | [feat] ITM Loss (#961) Summary: PR adds Image-text matching loss to MMF. The loss can be used standalone or with the MMFTransformer as a head. Addresses https://github.com/facebookresearch/mmf/issues/947 #466 - Adds ITM loss - Fixes ITM loss for validation dataset - Adds config for COCO ITM loss - Adds tests for ITM loss Pull Request resolved: https://github.com/facebookresearch/mmf/pull/961 Test Plan: Imported from GitHub, without a `Test Plan:` line. Pass tests Reviewed By: apsdehal Differential Revision: D28925014 Pulled By: vedanuj fbshipit-source-id: d123468f746820e1412380384dd15c5ebed15887 | 30 June 2021, 21:14:48 UTC |
refs/heads/gh/ytsheng/5/base | 273a56b | [feat] ITM Loss (#961) Summary: PR adds Image-text matching loss to MMF. The loss can be used standalone or with the MMFTransformer as a head. Addresses https://github.com/facebookresearch/mmf/issues/947 #466 - Adds ITM loss - Fixes ITM loss for validation dataset - Adds config for COCO ITM loss - Adds tests for ITM loss Pull Request resolved: https://github.com/facebookresearch/mmf/pull/961 Test Plan: Imported from GitHub, without a `Test Plan:` line. Pass tests Reviewed By: apsdehal Differential Revision: D28925014 Pulled By: vedanuj fbshipit-source-id: d123468f746820e1412380384dd15c5ebed15887 | 30 June 2021, 21:14:48 UTC |
refs/heads/gh/ytsheng/6/base | 273a56b | [feat] ITM Loss (#961) Summary: PR adds Image-text matching loss to MMF. The loss can be used standalone or with the MMFTransformer as a head. Addresses https://github.com/facebookresearch/mmf/issues/947 #466 - Adds ITM loss - Fixes ITM loss for validation dataset - Adds config for COCO ITM loss - Adds tests for ITM loss Pull Request resolved: https://github.com/facebookresearch/mmf/pull/961 Test Plan: Imported from GitHub, without a `Test Plan:` line. Pass tests Reviewed By: apsdehal Differential Revision: D28925014 Pulled By: vedanuj fbshipit-source-id: d123468f746820e1412380384dd15c5ebed15887 | 30 June 2021, 21:14:48 UTC |
refs/heads/gh/ytsheng/7/base | 273a56b | [feat] ITM Loss (#961) Summary: PR adds Image-text matching loss to MMF. The loss can be used standalone or with the MMFTransformer as a head. Addresses https://github.com/facebookresearch/mmf/issues/947 #466 - Adds ITM loss - Fixes ITM loss for validation dataset - Adds config for COCO ITM loss - Adds tests for ITM loss Pull Request resolved: https://github.com/facebookresearch/mmf/pull/961 Test Plan: Imported from GitHub, without a `Test Plan:` line. Pass tests Reviewed By: apsdehal Differential Revision: D28925014 Pulled By: vedanuj fbshipit-source-id: d123468f746820e1412380384dd15c5ebed15887 | 30 June 2021, 21:14:48 UTC |
refs/heads/gh/ytsheng/8/base | 273a56b | [feat] ITM Loss (#961) Summary: PR adds Image-text matching loss to MMF. The loss can be used standalone or with the MMFTransformer as a head. Addresses https://github.com/facebookresearch/mmf/issues/947 #466 - Adds ITM loss - Fixes ITM loss for validation dataset - Adds config for COCO ITM loss - Adds tests for ITM loss Pull Request resolved: https://github.com/facebookresearch/mmf/pull/961 Test Plan: Imported from GitHub, without a `Test Plan:` line. Pass tests Reviewed By: apsdehal Differential Revision: D28925014 Pulled By: vedanuj fbshipit-source-id: d123468f746820e1412380384dd15c5ebed15887 | 30 June 2021, 21:14:48 UTC |
refs/heads/hydra | 8ead66a | [feat] Hydra integration Attempt to test out hydra | 10 December 2020, 02:12:16 UTC |
refs/heads/itm | 73f6792 | [feat] ITM Loss | 06 June 2021, 20:07:19 UTC |
refs/heads/main | 02a55a3 | [staticdocs][mmf] automatic update for docusaurus-plugin-internaldocs-fb@1.7.0 Differential Revision: D42920157 fbshipit-source-id: f4958274017b9e880c38ebbd6d7ea9411546e08d | 02 February 2023, 10:46:34 UTC |
refs/heads/mask_key_mmft | 829ea29 | [feat] Support custom mask key in MMF Transformer Now, you can specify custom mask key for a modality using `mask_key` attribute. This also fixes an issue which caused MMFT to fail when text modality was not present. | 17 June 2021, 20:02:27 UTC |
refs/heads/mm_alignamet | e681e8d | [fix] Check for distributed is_available (#365) Summary: Fixes https://github.com/facebookresearch/mmf/issues/353 Pull Request resolved: https://github.com/facebookresearch/mmf/pull/365 Reviewed By: apsdehal Differential Revision: D22266850 Pulled By: vedanuj fbshipit-source-id: ed45963dae38c3157ee6eaa3cdcab2607cd5a1a9 | 29 June 2020, 17:59:18 UTC |
refs/heads/mmf_interactive | c42f208 | Initial Commit | 10 June 2020, 00:38:56 UTC |
refs/heads/mmft_output_key | 9b34e1f | [feat] MMFT MLP head output key; mix model losses and MMFLosses This will allow doing multitasking on single batch easily through different heads with different output_key per task Mixing allows usage of heads that returns losses directly with heads that don't and rely on MMFLoss Test Plan: Tests have been added | 04 May 2021, 17:34:28 UTC |
refs/heads/multitask_training | b5b01bc | Multi-Task Training with Stop on Plateu | 10 July 2020, 03:47:33 UTC |
refs/heads/notebooks | 6b6df04 | [fix] Update mmf_hm_example notebook - Clear registry state before run | 27 April 2021, 19:35:00 UTC |
refs/heads/pl_upgrade | 954a145 | [fix] oss and fbcode pytorch lightning version sync | 22 June 2021, 00:55:15 UTC |
refs/heads/project/cycle-consistency | 666b256 | Add code for cycle consistency paper | 07 June 2019, 22:49:34 UTC |
refs/heads/project/m4c | 28e48ce | update TextCaps bibtex | 01 April 2020, 19:21:30 UTC |
refs/heads/pt_19 | 1fbf93c | [chores] PT upgrade | 26 June 2021, 20:26:56 UTC |
refs/heads/skip_optimizer_update | c33bdc2 | skip optimizer update when nan loss | 08 September 2021, 04:52:49 UTC |
refs/heads/stable_mmft | 584e26d | [chores] Upgrade to pytorch 1.6 and torchvision 0.7 (#546) Summary: - Upgrading to pytorch 1.6, which comes with new features required for upcoming torchscript changes - Optimizer keys are no longer non-deterministic in PT1.6, check pytorch [PR](https://github.com/pytorch/pytorch/pull/37347 ). Remove `skip_keys` check from checkpoint optimizer tests. - Fixed nucleus sampling tests - Update docs to reflect support for pytorch 1.6 - Update CI tests, CircleCI and Actions to use pytorch 1.6, torchvision 0.7 Test Plan : - Tested with visual bert and mmbt models on Hateful Memes - Running internally at FB (runs on pytorch master) Pull Request resolved: https://github.com/facebookresearch/mmf/pull/546 Reviewed By: apsdehal Differential Revision: D23558059 Pulled By: vedanuj fbshipit-source-id: 3d82978276462a3f3dea564ac33581240e284d03 | 08 September 2020, 22:59:00 UTC |
refs/heads/v0.4 | 4321e54 | [fix] Fix checkpoint loading in for context and data parallel | 22 December 2019, 09:31:19 UTC |
refs/heads/vilbert_multimodal | 7026ab4 | [fix] Ckpt zoo load, restore and other issues (#311) Summary: - Explictly have an option for loading zoo in _load function - This fixes the issues with restore and other loading when resume_zoo is specified Pull Request resolved: https://github.com/facebookresearch/mmf/pull/311 Reviewed By: vedanuj Differential Revision: D21958031 Pulled By: apsdehal fbshipit-source-id: b13ea97c6edf75e3b60f91279337e6430daddc6f | 09 June 2020, 20:30:10 UTC |
refs/heads/vilbert_multitask | 77a14f4 | working code | 10 July 2020, 03:57:59 UTC |
refs/heads/vilbert_multitask_flickr30k | 48c0d58 | Working training | 01 July 2020, 17:58:35 UTC |
refs/heads/vilbert_multitask_visual7w | 0a1b540 | Addressing review comments | 10 July 2020, 18:19:49 UTC |
refs/heads/xla-checkpoint-fix | 54c8b11 | [fix] fix TPU checkpoint saving, parameter broadcasting, and misc stuff Fix the following on TPU training: - instead of using `xm.save`, add a custom `save_xla_ckpt` to avoid crashing on lr_scheduler state dict; **note that if we later switch back to xm.save, we need to set `global_master=True` in xm.save** - only delete old checkpoint from master process when using XLA - only open file in "wb" mode from the master process when using XLA. (Otherwise, concurrent writing to the same file in `with PathManager.open(ckpt_filepath, "wb")` from multiple processes causes permission denied error). - only save `config.yaml` from the global master process. - skip batch size validation in XLA mode in evaluation loop (since data loader already automatically drops the last batch in XLA mode) - On TPUs, there isn't an equivalent to DistributedDaraParallel yet, so adding a manual parameter broadcasting at the beginning of training (`broadcast_xla_master_model_param`). Otherwise we will have different parameters on each process since we are explicitly setting different seeds in different processes in MMF. - set `drop_last=True` in DistributedSampler for XLA to correctly drop the last batch in XLA mode (the current codebase only set `drop_last=True` in DataLoader, which is not enough). Test plan: tested on GCP. | 01 June 2021, 21:41:22 UTC |
refs/tags/v0.3 | c6796cc | [feat] Add script for feature extraction from vmb (#93) * [feat] Add script for feature extraction from vmb * [fix] Address comments in the PR * [fix] Address Meet's comments * [fix] Remove os.path.exists from download_file | 12 June 2019, 20:04:39 UTC |
refs/tags/v0.3.1 | 2df9160 | [fix] Fix detected objects in extract_features_vmb (#159) * [fix] Correctly ignore the background class * [fix] Output the correct number of objects | 22 August 2019, 21:47:29 UTC |