Skip to content

GitLab

Switch branch/tag
  1. 20 Oct, 2021 5 commits
    • Yuxin Wu's avatar
      print flow id in logs · 74a031b8
      Yuxin Wu authored 
      Summary: helps debugging

Reviewed By: zhanghang1989

Differential Revision: D31806396

fbshipit-source-id: 870308990c4c0c71453d107628b8adcb9edcf391
      74a031b8
    • Yanghan Wang's avatar
      toy example of training model for turing · ee9602a1
      Yanghan Wang authored 
      Summary:
Add toy example to illustrate the Turing workflow.
- modify the model building, add converting to helios step. Note that we need to hide this from OSS, so create FB version of the runner, in order to modify `build_model` and `get_default_cfg`.
- make the `D2 (https://github.com/facebookresearch/d2go/commit/7992f91324aee6ae59795063a007c6837e60cdb8)GoCompatibleMNISTRunner` up-to-date, and use the "tutorial" meta-arch for writing unit test since it's the simplest model. Note that even `TutorialNet` is very simple, there's still a constraint that the FC has to run on 4D tensor with 1x1 spatial dimension because it's been mapped to 1x1 Conv by Helios, modify the `TutorialNet` to make it compatible.

Reviewed By: newstzpz

Differential Revision: D31705305

fbshipit-source-id: 77949dfbf08252be5495e9273210274c8ad86abb
      ee9602a1
    • Yanghan Wang's avatar
      use fb. in import path - modeling/backbone/fb · 274d3b49
      Yanghan Wang authored 
      Summary: see bottom diff

Reviewed By: newstzpz

Differential Revision: D31780235

fbshipit-source-id: ec1285c4c5457a631e1eb88bebd47c9f41b47e12
      274d3b49
    • Peizhao Zhang's avatar
      Supported learnable qat. · f6ce583e
      Peizhao Zhang authored 
      Summary:
Supported learnable qat.
* Added a config key `QUANTIZATION.QAT.FAKE_QUANT_METHOD` to specify the qat metod (`default` or `learnable`).
* Added a config key `QUANTIZATION.QAT.ENABLE_LEARNABLE_OBSERVER_ITER` to specify the start iteration for learnable observers (before that it is using static observers).
* Custom quantization code needs to call ` d2go.utils.qat_utils.get_qat_qconfig()` to get proper qconfig for learnable qat. An exception will raise if qat method is learnable but no learnable observers are used in the model.
* Set the weight decay for scale/zero_point to 0 for the optimizer automatically.
* The way to use larnable qat: enable static observers -> enable fake quant -> enable learnable observers -> freeze bn.

Differential Revision: D31370822

fbshipit-source-id: a5a5044a539d0d7fe1cc6b36e6821fc411ce752a
      f6ce583e
    • Peizhao Zhang's avatar
      Refactored qat related code. · ef9c20cc
      Peizhao Zhang authored 
      Summary:
Refactored qat related code.
* Moved `_prepare_model_for_qat` related code to a function.
* Moved `_setup_non_qat_to_qat_state_dict_map` related code to a function.
* Moved QATHook related code to the quantization file and implemented as a class.

Differential Revision: D31370819

fbshipit-source-id: 836550b2c8d68cd93a84d5877ad9cef6f0f0eb39
      ef9c20cc
  3. 16 Oct, 2021 1 commit
  5. 15 Oct, 2021 2 commits
    • Peizhao Zhang's avatar
      Supported specifying customized parameter groups from model. · 87ce583c
      Peizhao Zhang authored 
      Summary:
Supported specifying customized parameter groups from model.
* Allow model to specify customized parameter groups by implementing a function `model.get_optimizer_param_groups(cfg)`
* Supported model with ddp.

Reviewed By: zhanghang1989

Differential Revision: D31289315

fbshipit-source-id: c91ba8014508e9fd5f172601b9c1c83c188338fd
      87ce583c
    • Peizhao Zhang's avatar
      Refactor for get_optimizer_param_groups. · 2dc3bc02
      Peizhao Zhang authored 
      Summary:
Refactor for get_optimizer_param_groups.
* Split `get_default_optimizer_params()` into multiple functions:
  * `get_optimizer_param_groups_default()`
  * `get_optimizer_param_groups_lr()`
  * `get_optimizer_param_groups_weight_decay()`
* Regroup the parameters to create the minimal amount of groups.
* Print all parameter groups when the optimizer is created.
    Param group 0: {amsgrad: False, betas: (0.9, 0.999), eps: 1e-08, lr: 10.0, params: 1, weight_decay: 1.0}
    Param group 1: {amsgrad: False, betas: (0.9, 0.999), eps: 1e-08, lr: 1.0, params: 1, weight_decay: 1.0}
    Param group 2: {amsgrad: False, betas: (0.9, 0.999), eps: 1e-08, lr: 1.0, params: 2, weight_decay: 0.0}
* Add some unit tests.

Reviewed By: zhanghang1989

Differential Revision: D31287783

fbshipit-source-id: e87df0ae0e67343bb2130db945d8faced44d7411
      2dc3bc02
  7. 14 Oct, 2021 1 commit
    • Yuxin Wu's avatar
      update benchmark_storage with instructions · 46f16a5e
      Yuxin Wu authored 
      Summary: Also modify launch() because it should not assume it's always called with a CfgNode object.

Differential Revision: D31494215

fbshipit-source-id: 8f07e9cb64969f8a14641956f7ef7c7160748bd9
      46f16a5e
  9. 13 Oct, 2021 2 commits
  11. 09 Oct, 2021 1 commit
    • Tao Xu's avatar
      fix real data driving generation in _generate() · 3b23dd39
      Tao Xu authored 
      Summary: Fix a failure bug in real image driving generating

Reviewed By: yc-fb

Differential Revision: D31362721

fbshipit-source-id: b222745aada1bd6680ca931d49a70d8b428828a6
      3b23dd39
  13. 07 Oct, 2021 2 commits
    • Yanghan Wang's avatar
      only evaluate EMA model on non-predictor models · d99428a1
      Yanghan Wang authored 
      Summary:
EMA is only applicable when testing non-predictor based models, this diff simply add a check so it won't evaluate ema models.

Side note: `do_test` should probably just handle single model, in the case of EMA, we could let `do_train` to return two models with and without ema, and call `do_test` on each of them. Then the temporary fix in this diff is not needed at all.

Reviewed By: wrlife

Differential Revision: D31450572

fbshipit-source-id: 8696922a9fd194f91315d2f3480dc8bfd8f36a3d
      d99428a1
    • Yuxin Wu's avatar
      remove SOLVER.STEPS from configs · 79ea94d5
      Yuxin Wu authored 
      Summary:
the LR scheduler is cosine, so this config has no effect.
Remove it to avoid confusion.

Reviewed By: sstsai-adl

Differential Revision: D31444047

fbshipit-source-id: b40e0d7d923c3b55dfe23353050ea0238b3afd16
      79ea94d5
  15. 06 Oct, 2021 1 commit
  17. 01 Oct, 2021 2 commits
  19. 27 Sep, 2021 2 commits
    • Yanghan Wang's avatar
      support scripting for torchscript ExportMethod · a9dce74e
      Yanghan Wang authored 
      Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/118

This diff adds the proper support for using scripting when exporting model.

Rename tracing-related code:
- Previously `trace_and_save_torchscript` is the primary function to export model, replace it with `export_optimize_and_save_torchscript`.
- Also rename `D2 (https://github.com/facebookresearch/d2go/commit/7992f91324aee6ae59795063a007c6837e60cdb8)TorchscriptTracingExport` to `TracingAdaptedTorchscriptExport` since it's not only for tracing now.

Introduce `jit_mode`:
- Add `jit_mode` option as the `export_kwargs` of ExportMethod.
- Add `scripting` and `tracing` trigger words to overwrite `jit_mode`. Please note that the `tracing` now applies to all models, which is different from the previous meaning (using `TracingAdapter` for RCNN).
- Therefore there're two ways of using scripting mode, 1) setting `jit_mode` in prepare_for_export; 2) using `script` trigger words. Add unit test as examples to illustrate two ways.
- Don't use `TracingAdapter` when scripting since it's not scriptable.

Consolidate triggering words logic.
- Group logic of handling trigger words (eg. `_mobile`, `_int8`, `scripting`, `tracing`) into a single decorator `update_export_kwargs_from_export_method` for better structuring and readability.

Reviewed By: zhanghang1989

Differential Revision: D31181624

fbshipit-source-id: 5fbb0d4fa4c29ffa4a761af8ea8f93b4bad4cef9
      a9dce74e
    • Yanghan Wang's avatar
      don't register @legacy as part of export method name · 8adb146e
      Yanghan Wang authored 
      Summary: Pull Request resolved: https://github.com/facebookresearch/d2go/pull/119

Reviewed By: zhanghang1989

Differential Revision: D31181216

fbshipit-source-id: 428116f4f4144e20410222825a9a00f75253ef4a
      8adb146e
  21. 24 Sep, 2021 5 commits
  23. 22 Sep, 2021 1 commit
    • Lei Tian's avatar
      fix optimizer setting in pytorch lightning · ea6e9f7f
      Lei Tian authored 
      Summary: fix optimizer setting in pytorch lightning

Reviewed By: wat3rBro

Differential Revision: D30988441

fbshipit-source-id: fcd2f4c77a87a790d7e99b0e3c833c291fd66e77
      ea6e9f7f
  25. 21 Sep, 2021 2 commits
  27. 20 Sep, 2021 2 commits
    • Yanghan Wang's avatar
      merge internal data build files · 07c4e54c
      Yanghan Wang authored 
      Reviewed By: ppwwyyxx

Differential Revision: D31035247

fbshipit-source-id: 7340e6f6bb813e284416e37060d0d511c5c79e03
      07c4e54c
    • Shiyu Dong's avatar
      Check if new_ds_name registered to MetadataCatalog before removing · f4fcff31
      Shiyu Dong authored 
      Summary:
As title, sometimes new_ds_name is not registered so it crashes the program when calling remove(). Adding a check.
A side effect to this is if it's not registered, get() method will register it first and then remove() will remove it from registery.

Reviewed By: ppwwyyxx

Differential Revision: D31049303

fbshipit-source-id: 149168fb89fd3b661b60717ff2aafa7a9bd52849
      f4fcff31
  29. 18 Sep, 2021 2 commits
  31. 15 Sep, 2021 2 commits
  33. 10 Sep, 2021 1 commit
  35. 09 Sep, 2021 1 commit
  37. 08 Sep, 2021 1 commit
  39. 02 Sep, 2021 2 commits
    • Lydia Chan's avatar
      Increase limit on number of detections per image in {COCO,LVIS}Evaluator · 2fb273ab
      Lydia Chan authored 
      Summary:
## Context
- The current limit on the number of detections per image (`K`) in LVIS is 300.
- Implementing AP_pool/AP_fixed requires removing this default limit on `K`
- [Literature](https://arxiv.org/pdf/2102.01066.pdf) has shown that increasing `K` correlates with AP gains

## This Diff
- Changed limit on number of detections per image (`K`) to be customizable for LVIS and COCO through `TEST.DETECTIONS_PER_IMAGE` in the config
   - For COCO:
       - Maintain the default `max_dets_per_image` to be [1, 10, 100] as from [COCOEval](https://www.internalfb.com/code/fbsource/[88bb57c3054a]/fbcode/deeplearning/projects/cocoApi/PythonAPI/pycocotools/cocoeval.py?lines=28-29)
       - Allow users to input a custom integer for `TEST.DETECTIONS_PER_IMAGE` in the config, and use  [1, 10, `TEST.DETECTIONS_PER_IMAGE`] for COCOEval
   - For LVIS:
       - Maintain the default `max_dets_per_image` to be 300 as from [LVISEval](https://www.internalfb.com/code/fbsource/[f6b86d023721]/fbcode/deeplearning/projects/lvisApi/lvis/eval.py?lines=528-529)
       - Allow users to input a custom integer for `TEST.DETECTIONS_PER_IMAGE` in the config, and use this in LVISEval
- Added `COCOevalMaxDets` for evaluating AP with the custom limit on number of detections per image (since default `COCOeval` uses 100 as limit on detections per image for evaluating AP)

## Inference Runs using this Diff
- Performed inference using `K = {300, 1000, 10000, 100000}`
- Launched fblearner flows for object detector baseline models with N1055536 (LVIS) and N1055756 (COCO)
  - Recorded [results of running inference](https://docs.google.com/spreadsheets/d/1rgdjN2KvxcYfKCkGUC4tMw0XQJ5oZL0dwjOIh84YRg8/edit?usp=sharing)

Reviewed By: ppwwyyxx

Differential Revision: D30077359

fbshipit-source-id: 372eb5e0d7c228fb77fe23bf80d53597ec66287b
      2fb273ab
    • Zhicheng Yan's avatar
      clamp reference point max to 1.0 to avoid NaN in regressed bbox · 0a38f8c8
      Zhicheng Yan authored 
      Summary:
For training DF-DETR with swin-transformer backbone which uses large size_divisibility 224 (=32 * 7) and potentially has more zero-padding, we find the regressed box can contain NaN values and fail the assertion here (https://fburl.com/code/p27ztcce).

This issue might be caused by two potential reasons.
- Fix 1. In DF-DETR encoder, the reference points prepared by `get_reference_points()` can contain normalized x,y coordinates larger than 1 due to the rounding issues during mask interpolation across feature scales (specific examples can be given upon request LoL). Thus, we clamp max of x,y coordinates to 1.0.

- Fix 2. The MLP used in bbox_embed heads contains 3 FC layers, which might be too many. We introduce an argument `BBOX_EMBED_NUM_LAYERS` to allow users to configure the number of FC layers. This change is back-compatible.

Reviewed By: zhanghang1989

Differential Revision: D30661167

fbshipit-source-id: c7e94983bf1ec07426fdf1b9d363e5163637f21a
      0a38f8c8
  41. 31 Aug, 2021 2 commits